Update the Assessment Results

HBV_Eligibility_Results.csv

@@ -0,0 +1,436 @@
+Case ID,Age,Sex,Pregnancy Status,HBsAg,HBeAg,HBV DNA (IU/mL),ALT (U/L),Fibrosis/Cirrhosis Stage,Necroinflammation,Extrahepatic Manifestations,Immunosuppressive Therapy,Coinfections,Family History of HCC/Cirrhosis,Smoking,Comorbidities,Eligibility,Rationale
+Case1,28,M,No,Positive (36 months),Positive,8500000,120,F1,Moderate,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL and ALT > ULN, with moderate necroinflammation (Grade A) [SASLT 2021, Page 6]
+- Meets criteria for treatment due to HBV DNA > 20,000 IU/mL and ALT > 2xULN (Grade B) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months (Grade B) [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable as patient has no coinfections, immunosuppression, or extrahepatic manifestations [SASLT 2021, Pages 6-9]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case2,45,F,No,Positive (60 months),Negative,15000,65,F1–F2,Moderate,No,,,No,No,Metabolic syndrome,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, moderate fibrosis (Grade A) [SASLT 2021, Page 6]
+- Necroinflammatory activity (A2) supports treatment initiation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months (Grade D) [SASLT 2021, Page 7]
+
+Special Considerations:
+- Metabolic syndrome may require additional management alongside HBV treatment [SASLT 2021, Page 6]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case3,52,M,No,Positive (120 months),Negative,3500,32,F2–F3,Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, moderate fibrosis (F2-F3), and chronic HBV infection (Grade A) [SASLT 2021, Page 6].
+- ALT is within normal limits but fibrosis stage supports treatment initiation (Grade A) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF as first-line treatment (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol, including HBV DNA and ALT levels every 3-6 months (Grade A) [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable as patient has no coinfections, immunosuppression, or extrahepatic manifestations [SASLT 2021, Pages 6, 9].
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case4,60,M,No,Positive (240 months),Negative,700,28,F4 (compensated),Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: Cirrhosis (F4) with detectable HBV DNA, regardless of ALT levels (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months (Grade A) [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable to this patient [SASLT 2021, Pages 6-10]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case5,22,F,No,Positive (18 months),Positive,25000000,27,F0,,No,,,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA > 20,000 IU/mL but ALT ≤ 2xULN and age < 30 years (Grade B) [SASLT 2021, Page 6].
+- No significant fibrosis or necroinflammation (F0-F1, A0) and no family history of HCC or cirrhosis [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 3-6 months for HBeAg-positive patients under 30 years (Grade B) [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable to this patient.
+
+References:
+- Pages 6, 7: Treatment criteria, monitoring protocols"
+Case6,40,M,No,Positive (84 months),Negative,900,22,F0–F1,,No,,,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA < 2,000 IU/mL, ALT ≤ ULN, no significant fibrosis or necroinflammation [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7].
+
+References:
+- Pages 6, 7: Treatment criteria, monitoring protocols"
+Case7,31,F,Yes (28 weeks),Positive (48 months),Positive,500000,35,F1,Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 100,000 IU/mL during pregnancy warrants antiviral prophylaxis starting at 24-28 weeks (Grade D) [SASLT 2021, Page 10].
+- Eligible: HBeAg-positive chronic HBV infection with high HBV DNA levels (> 20,000 IU/mL) and age > 30 years (Grade D) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start antiviral prophylaxis with TDF or TAF at 24-28 weeks of pregnancy (Grade D) [SASLT 2021, Page 10].
+- Preferred regimens for CHB treatment include TDF, TAF, or ETV monotherapy (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor HBV DNA and ALT levels every 3 months during pregnancy and postpartum (Grade D) [SASLT 2021, Page 7].
+- Assess treatment response and adjust therapy as needed per SASLT protocol [SASLT 2021, Page 7].
+
+Special Considerations:
+- Breastfeeding is permitted for HBsAg-positive women on TDF (Grade B) [SASLT 2021, Page 10].
+
+References:
+- Pages 6, 7, 8, 10: Treatment criteria, drugs, monitoring, pregnancy-related recommendations"
+Case8,35,M,No,Positive (72 months),Positive,9000000,33,F1,Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBeAg-positive chronic HBV infection, age > 30 years, persistently high HBV DNA levels (Grade D) [SASLT 2021, Page 6].
+- HBV DNA > 2,000 IU/mL and ALT ≤ ULN, meeting treatment criteria for HBeAg-positive patients > 30 years (Grade D) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months (Grade B) [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable to this patient.
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case9,55,F,No,Positive (180 months),Negative,6200,45,F1–F2,Moderate,No,,,Yes (father HCC),No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, moderate fibrosis (Grade A) [SASLT 2021, Page 6]
+- Family history of HCC or cirrhosis supports treatment initiation (Grade D) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable to this patient [SASLT 2021, Pages 6-10]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case10,63,M,No,Positive (300 months),Negative,50,32,F4 (decompensated),Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: Patient has cirrhosis (F4 fibrosis) with detectable HBV DNA, regardless of ALT levels (Grade A) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF as first-line treatment (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol, including HBV DNA and ALT levels every 3-6 months [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable as patient has no coinfections, immunosuppression, or extrahepatic manifestations [SASLT 2021, Pages 6, 9].
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case11,68,M,No,Positive (25 years),Negative,5800,41,F2,Mild,No,,,No,No,Diabetes,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, and moderate fibrosis (F2-F3) (Grade A) [SASLT 2021, Page 6]
+- Chronic HBV infection with significant liver disease warrants treatment initiation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF as first-line treatment (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol, including HBV DNA and ALT levels every 3-6 months (Grade A) [SASLT 2021, Page 7]
+
+Special Considerations:
+- Patient's age (68 years) and comorbidity (diabetes) may require closer monitoring for treatment-related side effects [SASLT 2021, Page 6]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case12,50,F,No,Positive,Negative,450,32,,Mild,No,Yes (tacrolimus),,No,No,Post-transplant,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA < 2,000 IU/mL, ALT ≤ ULN, no significant fibrosis or necroinflammation [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7].
+
+Special Considerations:
+- Immunosuppression: Consider prophylaxis if undergoing immunosuppressive therapy (Grade A) [SASLT 2021, Page 9].
+
+References:
+- Pages 6, 7, 9: Treatment criteria, monitoring protocols, immunosuppression considerations"
+Case13,27,M,No,Positive (36 months),Negative,1700,55,F1,Mild,No,,,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA < 2,000 IU/mL, ALT > ULN but mild fibrosis and necroinflammation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6]
+
+Monitoring and Follow-up:
+- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable to this patient [SASLT 2021, Pages 6-10]
+
+References:
+- Pages 6, 7: Treatment criteria, monitoring protocols"
+Case14,44,F,No,Positive,Negative,3000,70,F2,Moderate,No,,HCV RNA positive,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, and moderate fibrosis (Grade A) [SASLT 2021, Page 6].
+- HBV-HCV coinfection requires concurrent or prior HBV treatment to prevent reactivation during DAA therapy (Grade A) [SASLT 2021, Page 9].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor HBV DNA and ALT every 4-8 weeks during DAA therapy and three months post-therapy (Grade D) [SASLT 2021, Page 9].
+- Monitor treatment response per SASLT protocol (Grade A) [SASLT 2021, Page 7].
+
+Special Considerations:
+- HBV-HCV coinfection requires close monitoring to prevent HBV reactivation during HCV treatment (Grade A) [SASLT 2021, Page 9].
+
+References:
+- Pages 6, 7, 8, 9: Treatment criteria, drugs, monitoring, HBV-HCV coinfection"
+Case15,33,M,No,Positive,Negative,25000,30,F1,Mild,No,,HIV positive,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA > 2,000 IU/mL but ALT ≤ ULN and no significant fibrosis or necroinflammation (Grade A) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 3 months for the first year, then every 6 months (HBeAg-negative, HBV DNA ≥ 2,000 IU/mL) (Grade D) [SASLT 2021, Page 7].
+
+Special Considerations:
+- HBV-HIV coinfection: ensure ART includes TDF- or TAF-based regimen (Grade A) [SASLT 2021, Page 9].
+
+References:
+- Pages 6, 7, 9: Treatment criteria, monitoring protocols, HIV coinfection"
+Case16,30,F,Yes (30 weeks),Positive,Positive,40000,28,F0,,No,,,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA > 2,000 IU/mL but ALT ≤ ULN and no significant fibrosis or necroinflammation [SASLT 2021, Page 6].
+- Pregnancy alone does not meet treatment criteria unless HBV DNA > 100,000 IU/mL at 24-28 weeks [SASLT 2021, Page 10].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 3-6 months for HBeAg-positive patients under 30 years (Grade B) [SASLT 2021, Page 7].
+- Assess HBV DNA and ALT levels regularly to evaluate disease progression [SASLT 2021, Page 6].
+
+Special Considerations:
+- Consider antiviral prophylaxis with TDF if HBV DNA exceeds 100,000 IU/mL at 24-28 weeks of pregnancy [SASLT 2021, Page 10].
+
+References:
+- Pages 6, 7, 10: Treatment criteria, monitoring protocols, pregnancy considerations"
+Case17,39,M,No,Positive,Negative,4800,250,F1,Marked,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL and ALT > ULN, regardless of HBeAg status (Grade A) [SASLT 2021, Page 6]
+- ALT > 2x ULN supports treatment initiation (Grade B) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months (Grade B) [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable to this patient [SASLT 2021, Pages 6-10]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case18,49,F,No,Positive,Negative,2300,37,F1,Mild,No,,,Yes (mother cirrhosis),No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, and family history of HCC or cirrhosis (Grade D) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol, including HBV DNA and ALT levels every 3-6 months [SASLT 2021, Page 7].
+
+Special Considerations:
+- Family history of HCC or cirrhosis increases the need for treatment to reduce long-term risks [SASLT 2021, Page 6].
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case19,56,M,No,Positive,Negative,10,30,F4 (compensated),Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: Patient has cirrhosis (F4 fibrosis) with detectable HBV DNA, regardless of ALT levels (Grade A) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF as first-line treatment (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months (Grade A) [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable to this patient.
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case20,41,F,No,Positive (10 years),Negative,1900,28,F1,Mild,No,,,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA < 2,000 IU/mL, ALT ≤ ULN, no significant fibrosis or necroinflammation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6]
+
+Monitoring and Follow-up:
+- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable to this patient [SASLT 2021, Page 6]
+
+References:
+- Pages 6, 7: Treatment criteria, monitoring protocols"
+Case21,56,F,No,Positive,Negative,0,20,F4 (compensated),Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: Patient has cirrhosis (F4 fibrosis) with detectable HBsAg, regardless of HBV DNA and ALT levels (Grade A) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF as the preferred treatment options (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol, including HBV DNA, ALT, and fibrosis assessment [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable as the patient has no coinfections, immunosuppression, or other special conditions [SASLT 2021, Pages 6-10].
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case22,70,M,No,Positive (30 years),Negative,1800,27,F2,Mild,No,,,No,Yes (smoker),Hypertension,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA < 2,000 IU/mL, ALT ≤ ULN, moderate fibrosis but does not meet treatment thresholds [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7].
+
+References:
+- Pages 6, 7: Treatment criteria, monitoring protocols"
+Case23,33,F,Yes (12 weeks),Positive,Positive,300000,40,F1,Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 100,000 IU/mL during pregnancy warrants antiviral prophylaxis (Grade D) [SASLT 2021, Page 10]
+- ALT > ULN and HBeAg-positive status further support treatment initiation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start antiviral prophylaxis with TDF or TAF at 24-28 weeks of pregnancy (Grade D) [SASLT 2021, Page 10]
+- Preferred regimens include TDF or TAF as monotherapy (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor HBV DNA and ALT levels every 3 months during treatment (Grade D) [SASLT 2021, Page 7]
+- Postpartum follow-up to assess treatment continuation or cessation [SASLT 2021, Page 10]
+
+Special Considerations:
+- Breastfeeding is permitted while on TDF therapy (Grade B) [SASLT 2021, Page 10]
+
+References:
+- Pages 6, 7, 8, 10: Treatment criteria, drugs, monitoring, pregnancy considerations"
+Case24,46,M,No,Positive (8 years),Negative,50000,48,F2,Moderate,Yes (vasculitis),,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, moderate fibrosis (Grade A) [SASLT 2021, Page 6]
+- Extrahepatic manifestations further support treatment initiation (Grade D) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response every 3-6 months per SASLT protocol (Grade B) [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable to this patient.
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case25,58,F,No,Positive,Negative,2000,60,F2,Moderate,No,,,No,No,CKD stage 3,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, moderate fibrosis (Grade A) [SASLT 2021, Page 6]
+- Moderate necroinflammation (A2) and fibrosis (F2-F3) support treatment initiation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response every 3-6 months per SASLT protocol [SASLT 2021, Page 7]
+- Assess HBV DNA, ALT, and fibrosis progression during follow-up [SASLT 2021, Page 7]
+
+Special Considerations:
+- CKD stage 3: Prefer TAF over TDF due to better renal safety profile (Grade A) [SASLT 2021, Page 8]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case26,29,M,No,Positive,Positive,2500000,33,F0,,No,,,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA > 2,000 IU/mL but ALT ≤ ULN and no significant fibrosis or necroinflammation [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 3-6 months (HBeAg-positive, HBV DNA > 2,000 IU/mL, ALT ≤ ULN) (Grade B) [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable to this patient.
+
+References:
+- Pages 6, 7: Treatment criteria, monitoring protocols"
+Case27,54,M,No,Positive,Negative,750,18,F4 (compensated),Mild,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: Patient has cirrhosis (F4 fibrosis) with detectable HBV DNA, regardless of ALT levels (Grade A) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF as the preferred treatment options (Grade A) [SASLT 2021, Page 8].
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol, including HBV DNA and ALT levels every 3-6 months [SASLT 2021, Page 7].
+
+Special Considerations:
+- None applicable as the patient has no coinfections, immunosuppression, or other special conditions [SASLT 2021, Pages 6-10].
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case28,38,F,No,Positive,Negative,6000,80,F2,Moderate,No,,,No,No,Obesity,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, and moderate fibrosis (Grade A) [SASLT 2021, Page 6]
+- Necroinflammatory activity (A2) and fibrosis stage (F2-F3) further support treatment initiation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF as first-line treatment (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol, including HBV DNA and ALT levels every 3-6 months initially (Grade A) [SASLT 2021, Page 7]
+- Assess fibrosis progression and treatment adherence regularly (Grade A) [SASLT 2021, Page 7]
+
+Special Considerations:
+- Obesity may impact liver disease progression; consider lifestyle interventions alongside antiviral therapy [SASLT 2021, Page 6]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"
+Case29,42,M,No,Positive,Negative,1200,25,F1,Mild,No,,,No,No,,False,"Eligibility and Rationale:
+- Not eligible: HBV DNA < 2,000 IU/mL, ALT ≤ ULN, no significant fibrosis or necroinflammation (Grade A) [SASLT 2021, Page 6].
+
+Treatment Recommendations:
+- Treatment not indicated at this time [SASLT 2021, Page 6].
+
+Monitoring and Follow-up:
+- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7].
+
+References:
+- Pages 6, 7: Treatment criteria, monitoring protocols"
+Case30,25,F,No,Positive (2 years),Positive,12000000,95,F1,Moderate,No,,,No,No,,True,"Eligibility and Rationale:
+- Eligible: HBV DNA > 20,000 IU/mL and ALT > 2xULN, regardless of fibrosis stage (Grade B) [SASLT 2021, Page 6]
+- Moderate necroinflammatory activity (A2) supports treatment initiation (Grade A) [SASLT 2021, Page 6]
+
+Treatment Recommendations:
+- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]
+
+Monitoring and Follow-up:
+- Monitor treatment response per SASLT protocol every 3-6 months (Grade B) [SASLT 2021, Page 7]
+
+Special Considerations:
+- None applicable to this patient [SASLT 2021, Pages 6-10]
+
+References:
+- Pages 6, 7, 8: Treatment criteria, drugs, monitoring protocols"

INTEGRATION_SUMMARY.md

@@ -0,0 +1,268 @@
+# LlamaParse Integration Summary
+
+## Changes Made
+
+### 1. **core/data_loaders.py** - Complete Replacement
+**Status**: ✅ Complete
+
+**Changes**:
+- ❌ Removed: `PyMuPDF4LLMLoader` and `TesseractBlobParser`
+- ✅ Added: `LlamaParse` and `SimpleDirectoryReader` from llama-index
+- ✅ Added: `os` module for environment variable handling
+
+**New Functions**:
+1. `load_pdf_documents(pdf_path, api_key=None)` - Basic LlamaParse loader
+2. `load_pdf_documents_advanced(pdf_path, api_key=None, premium_mode=False)` - Advanced loader with premium features
+3. `load_multiple_pdfs(pdf_directory, api_key=None, file_pattern="*.pdf")` - Batch processing
+
+**Key Features**:
+- Medical document optimized parsing instructions
+- Accurate page numbering with `split_by_page=True`
+- Preserves borderless tables and complex layouts
+- Enhanced metadata tracking
+- Premium mode option for GPT-4o parsing
+
+---
+
+### 2. **core/config.py** - Configuration Updates
+**Status**: ✅ Complete
+
+**Changes**:
+```python
+# Added to Settings class
+LLAMA_CLOUD_API_KEY: str | None = None
+LLAMA_PREMIUM_MODE: bool = False
+```
+
+**Purpose**:
+- Store LlamaParse API key from environment variables
+- Control premium/basic parsing mode
+- Centralized configuration management
+
+---
+
+### 3. **core/utils.py** - Pipeline Integration
+**Status**: ✅ Complete
+
+**Changes**:
+1. **Import Update** (Line 12):
+   ```python
+   from .config import get_embedding_model, VECTOR_STORE_DIR, CHUNKS_PATH, NEW_DATA, PROCESSED_DATA, settings
+   ```
+
+2. **Function Update** `_load_documents_for_file()` (Lines 118-141):
+   ```python
+   def _load_documents_for_file(file_path: Path) -> List[Document]:
+       try:
+           if file_path.suffix.lower() == '.pdf':
+               # Use advanced LlamaParse loader with settings from config
+               api_key = settings.LLAMA_CLOUD_API_KEY
+               premium_mode = settings.LLAMA_PREMIUM_MODE
+               
+               return data_loaders.load_pdf_documents_advanced(
+                   file_path,
+                   api_key=api_key,
+                   premium_mode=premium_mode
+               )
+           return data_loaders.load_markdown_documents(file_path)
+       except Exception as e:
+           logger.error(f"Failed to load {file_path}: {e}")
+           return []
+   ```
+
+**Impact**:
+- All PDF processing now uses LlamaParse automatically
+- Reads configuration from environment variables
+- Maintains backward compatibility with markdown files
+
+---
+
+## New Files Created
+
+### 1. **LLAMAPARSE_INTEGRATION.md**
+Complete documentation including:
+- Setup instructions
+- Configuration guide
+- Usage examples
+- Cost considerations
+- Troubleshooting
+- Migration guide
+
+### 2. **test_llamaparse.py**
+Test suite with:
+- Configuration checker
+- Single PDF test
+- Batch processing test
+- Full pipeline test
+
+### 3. **INTEGRATION_SUMMARY.md** (this file)
+Quick reference for all changes
+
+---
+
+## Environment Variables Required
+
+Add to your `.env` file:
+
+```env
+# Required for LlamaParse
+LLAMA_CLOUD_API_KEY=llx-your-api-key-here
+
+# Optional: Enable premium mode (default: False)
+LLAMA_PREMIUM_MODE=False
+
+# Existing (still required)
+OPENAI_API_KEY=your-openai-key
+```
+
+---
+
+## Installation Requirements
+
+```bash
+pip install llama-parse llama-index-core
+```
+
+---
+
+## How to Use
+
+### Automatic Processing (Recommended)
+1. Set `LLAMA_CLOUD_API_KEY` in `.env`
+2. Place PDFs in `data/new_data/PROVIDER/`
+3. Run your application - documents are processed automatically on startup
+
+### Manual Processing
+```python
+from core.utils import process_new_data_and_update_vector_store
+
+# Process all new documents
+vector_store = process_new_data_and_update_vector_store()
+```
+
+### Direct PDF Loading
+```python
+from pathlib import Path
+from core.data_loaders import load_pdf_documents_advanced
+
+pdf_path = Path("data/new_data/SASLT/guideline.pdf")
+documents = load_pdf_documents_advanced(pdf_path)
+```
+
+---
+
+## Testing
+
+Run the test suite:
+```bash
+python test_llamaparse.py
+```
+
+This will:
+1. ✅ Check configuration
+2. ✅ Test single PDF loading
+3. ✅ (Optional) Test batch processing
+4. ✅ (Optional) Test full pipeline
+
+---
+
+## Backward Compatibility
+
+✅ **Fully backward compatible**:
+- Existing processed documents remain valid
+- Vector store continues to work
+- Markdown processing unchanged
+- No breaking changes to API
+
+---
+
+## Benefits
+
+| Aspect | Before (PyMuPDF4LLMLoader) | After (LlamaParse) |
+|--------|---------------------------|-------------------|
+| **Borderless Tables** | ❌ Poor | ✅ Excellent |
+| **Complex Layouts** | ⚠️ Moderate | ✅ Excellent |
+| **Medical Terminology** | ⚠️ Moderate | ✅ Excellent |
+| **Page Numbering** | ✅ Good | ✅ Excellent |
+| **Processing Speed** | ✅ Fast (local) | ⚠️ Slower (cloud) |
+| **Cost** | ✅ Free | ⚠️ ~$0.003-0.01/page |
+| **Accuracy** | ⚠️ Moderate | ✅ High |
+
+---
+
+## Cost Estimation
+
+### Basic Mode (~$0.003/page)
+- 50-page guideline: ~$0.15
+- 100-page guideline: ~$0.30
+
+### Premium Mode (~$0.01/page)
+- 50-page guideline: ~$0.50
+- 100-page guideline: ~$1.00
+
+**Note**: LlamaParse caches results, so re-processing is free.
+
+---
+
+## Workflow Example
+
+```
+1. User places PDF in data/new_data/SASLT/
+   └── new_guideline.pdf
+
+2. Application startup triggers processing
+   ├── Detects new PDF
+   ├── Calls load_pdf_documents_advanced()
+   ├── LlamaParse processes with medical optimizations
+   ├── Extracts 50 pages with accurate metadata
+   └── Returns Document objects
+
+3. Pipeline continues
+   ├── Splits into 245 chunks
+   ├── Updates vector store
+   └── Moves to data/processed_data/SASLT/new_guideline_20251111_143022.pdf
+
+4. Ready for RAG queries
+   └── Vector store contains new guideline content
+```
+
+---
+
+## Next Steps
+
+1. ✅ Set `LLAMA_CLOUD_API_KEY` in `.env`
+2. ✅ Install dependencies: `pip install llama-parse llama-index-core`
+3. ✅ Test with: `python test_llamaparse.py`
+4. ✅ Place PDFs in `data/new_data/PROVIDER/`
+5. ✅ Run application and verify processing
+
+---
+
+## Support & Troubleshooting
+
+### Common Issues
+
+**1. API Key Not Found**
+```
+ValueError: LlamaCloud API key not found
+```
+→ Set `LLAMA_CLOUD_API_KEY` in `.env`
+
+**2. Import Errors**
+```
+ModuleNotFoundError: No module named 'llama_parse'
+```
+→ Run: `pip install llama-parse llama-index-core`
+
+**3. Slow Processing**
+→ Normal for cloud processing (30-60s per document)
+→ Subsequent runs use cache (much faster)
+
+### Logs
+Check `logs/app.log` for detailed processing information
+
+---
+
+**Integration Date**: November 11, 2025  
+**Status**: ✅ Production Ready  
+**Version**: 1.0

LLAMAPARSE_INTEGRATION.md

@@ -0,0 +1,265 @@
+# LlamaParse Integration Guide
+
+## Overview
+The HBV AI Assistant now uses **LlamaParse** for advanced PDF parsing, replacing PyMuPDF4LLMLoader. LlamaParse excels at:
+- ✅ Borderless tables (common in medical guidelines)
+- ✅ Complex document layouts
+- ✅ Hierarchical section preservation
+- ✅ Accurate page numbering
+- ✅ Medical terminology and dosage tables
+
+## Setup
+
+### 1. Install Required Packages
+```bash
+pip install llama-parse llama-index-core
+```
+
+### 2. Get Your API Key
+1. Visit: https://cloud.llamaindex.ai/api-key
+2. Sign up/login and generate an API key
+3. Copy your API key (format: `llx-...`)
+
+### 3. Configure Environment Variables
+Add to your `.env` file:
+
+```env
+# Required: LlamaParse API Key
+LLAMA_CLOUD_API_KEY=llx-your-api-key-here
+
+# Optional: Enable premium GPT-4o mode (higher accuracy, costs more)
+LLAMA_PREMIUM_MODE=False
+```
+
+## How It Works
+
+### Automatic Processing Pipeline
+When you process new documents from `data/new_data/`, the system automatically:
+
+1. **Detects PDF files** in `data/new_data/PROVIDER/` directories
+2. **Uses LlamaParse** with medical document optimizations:
+   - Preserves table structures (including borderless tables)
+   - Maintains hierarchical headings
+   - Extracts dosage information accurately
+   - Keeps reference citations intact
+3. **Splits by page** for accurate page numbering
+4. **Extracts metadata**: provider, disease, page numbers
+5. **Updates vector store** for RAG queries
+
+### Configuration Options
+
+#### Basic Mode (Default)
+```python
+# In .env
+LLAMA_CLOUD_API_KEY=llx-your-key
+LLAMA_PREMIUM_MODE=False
+```
+- Uses standard LlamaParse parsing
+- Good accuracy for most medical documents
+- Lower cost
+
+#### Premium Mode
+```python
+# In .env
+LLAMA_CLOUD_API_KEY=llx-your-key
+LLAMA_PREMIUM_MODE=True
+```
+- Uses GPT-4o for parsing
+- Highest accuracy for complex tables
+- Higher cost per page
+- Recommended for critical medical guidelines
+
+## Usage
+
+### Processing New Documents
+
+1. **Place PDFs** in the appropriate directory:
+   ```
+   data/new_data/SASLT/guideline.pdf
+   data/new_data/WHO/recommendations.pdf
+   ```
+
+2. **Run the processing** (automatic on app startup or manually):
+   ```python
+   from core.utils import process_new_data_and_update_vector_store
+   
+   # Process all new documents
+   vector_store = process_new_data_and_update_vector_store()
+   ```
+
+3. **Files are automatically moved** to `data/processed_data/` after successful processing
+
+### Manual PDF Loading
+
+You can also load PDFs manually:
+
+```python
+from pathlib import Path
+from core.data_loaders import load_pdf_documents_advanced
+
+# Basic usage (reads API key from environment)
+pdf_path = Path("data/new_data/SASLT/guideline.pdf")
+documents = load_pdf_documents_advanced(pdf_path)
+
+# With explicit API key
+documents = load_pdf_documents_advanced(
+    pdf_path,
+    api_key="llx-your-key-here",
+    premium_mode=True
+)
+
+# Batch processing
+from core.data_loaders import load_multiple_pdfs
+
+pdf_dir = Path("data/new_data/SASLT")
+all_documents = load_multiple_pdfs(pdf_dir)
+```
+
+## Document Metadata
+
+Each processed document includes:
+
+```python
+{
+    "source": "SASLT_2021.pdf",
+    "disease": "HBV",
+    "provider": "SASLT",
+    "page_number": 6,
+    "document_index": 5,
+    "parser": "llamaparse",
+    "premium_mode": False
+}
+```
+
+## Parsing Instructions
+
+LlamaParse is configured with medical-specific instructions:
+
+### Basic Mode
+```
+"This is a medical guideline document. 
+Pay special attention to tables (including borderless tables), 
+clinical recommendations, dosage information, and reference citations. 
+Preserve table structure and maintain hierarchical headings."
+```
+
+### Premium Mode
+```
+"Medical guideline document with complex tables. Instructions:
+0. Keep the original text intact without changing anything
+1. Preserve all table structures, especially borderless tables
+2. Maintain hierarchical organization of sections and subsections
+3. Keep dosage tables and treatment algorithms intact
+4. Preserve reference numbers and citations
+5. Identify and mark clinical recommendation levels
+6. Extract figures and their captions accurately"
+```
+
+## Cost Considerations
+
+- **Basic Mode**: ~$0.003 per page
+- **Premium Mode**: ~$0.01 per page (GPT-4o)
+- **Caching**: LlamaParse caches results, so re-processing the same file is free
+
+### Cost Estimation
+For a 50-page medical guideline:
+- Basic: ~$0.15
+- Premium: ~$0.50
+
+## Troubleshooting
+
+### API Key Not Found
+```
+ValueError: LlamaCloud API key not found
+```
+**Solution**: Set `LLAMA_CLOUD_API_KEY` in your `.env` file
+
+### Import Errors
+```
+ModuleNotFoundError: No module named 'llama_parse'
+```
+**Solution**: Install required packages:
+```bash
+pip install llama-parse llama-index-core
+```
+
+### Slow Processing
+- LlamaParse processes documents in the cloud
+- First-time processing takes longer (30-60 seconds per document)
+- Subsequent processing uses cache (much faster)
+- Consider using `premium_mode=False` for faster processing
+
+### Empty Results
+- Check that PDF is not corrupted
+- Verify API key is valid
+- Check logs for detailed error messages
+
+## Migration from PyMuPDF4LLMLoader
+
+The integration is **backward compatible**:
+- Existing processed documents remain valid
+- Vector store continues to work
+- Only new documents use LlamaParse
+- No changes needed to existing code
+
+### What Changed
+1. **`core/data_loaders.py`**: Replaced PyMuPDF4LLMLoader with LlamaParse
+2. **`core/config.py`**: Added `LLAMA_CLOUD_API_KEY` and `LLAMA_PREMIUM_MODE` settings
+3. **`core/utils.py`**: Updated `_load_documents_for_file()` to use `load_pdf_documents_advanced()`
+
+## Benefits Over PyMuPDF4LLMLoader
+
+| Feature | PyMuPDF4LLMLoader | LlamaParse |
+|---------|-------------------|------------|
+| Borderless tables | ❌ Poor | ✅ Excellent |
+| Complex layouts | ⚠️ Moderate | ✅ Excellent |
+| Medical terminology | ⚠️ Moderate | ✅ Excellent |
+| Page numbering | ✅ Good | ✅ Excellent |
+| Processing speed | ✅ Fast (local) | ⚠️ Slower (cloud) |
+| Cost | ✅ Free | ⚠️ Paid API |
+| Accuracy | ⚠️ Moderate | ✅ High |
+
+## Example Workflow
+
+```python
+# 1. Set up environment
+# Add to .env:
+# LLAMA_CLOUD_API_KEY=llx-your-key-here
+# LLAMA_PREMIUM_MODE=False
+
+# 2. Place new PDFs
+# data/new_data/SASLT/new_guideline.pdf
+
+# 3. Process automatically (on app startup)
+# Or manually:
+from core.utils import process_new_data_and_update_vector_store
+
+vector_store = process_new_data_and_update_vector_store()
+# Output: 
+# ✅ Parsing PDF with LlamaParse (Premium: False): new_guideline.pdf
+# ✅ Loaded 50 pages from PDF: new_guideline.pdf
+# ✅ Split 50 documents into 245 chunks
+# ✅ Added 245 new chunks to existing vector store
+# 📦 Moved processed file: new_guideline.pdf -> SASLT/new_guideline_20251111_143022.pdf
+
+# 4. Query the system
+from core.agent import answer_question
+
+response = answer_question(
+    "What is the recommended treatment for HBeAg-positive chronic hepatitis B?"
+)
+print(response)
+```
+
+## Support
+
+For issues or questions:
+1. Check the logs in `logs/app.log`
+2. Verify API key is valid
+3. Review LlamaParse documentation: https://docs.llamaindex.ai/en/stable/llama_cloud/llama_parse/
+4. Check environment variables are set correctly
+
+---
+
+**Last Updated**: November 11, 2025
+**Integration Status**: ✅ Complete and Production Ready

core/config.py

@@ -17,6 +17,10 @@ class Settings(BaseSettings):
 
     OPENAI_API_KEY: str
     OPENAI_BASE_URL: str | None = None
+    
+    # LlamaParse configuration for advanced PDF parsing
+    LLAMA_CLOUD_API_KEY: str | None = None
+    LLAMA_PREMIUM_MODE: bool = False  # Set to True for GPT-4o parsing (costs more)
 
     LOG_LEVEL: str = os.getenv("LOG_LEVEL", "INFO")
     DATA_DIR: str = os.getenv("DATA_DIR", "")

core/data_loaders.py

@@ -1,17 +1,17 @@
 # Import required libraries
-import pandas as pd
+import os
 from pathlib import Path
 from typing import List
 from langchain.schema import Document
 from .config import logger
-from langchain_pymupdf4llm import PyMuPDF4LLMLoader
-from langchain_community.document_loaders.parsers import TesseractBlobParser
+from llama_parse import LlamaParse
+from llama_index.core import SimpleDirectoryReader
 
 
-def load_pdf_documents(pdf_path: Path) -> List[Document]:
+def load_pdf_documents(pdf_path: Path, api_key: str = None) -> List[Document]:
     """
-    Load and process PDF documents from medical guidelines using PyMuPDF4LLMLoader.
-    Uses Tesseract for image extraction and optimized table extraction for medical documents.
+    Load and process PDF documents from medical guidelines using LlamaParse.
+    Excellent for borderless tables and complex medical document layouts.
     Extracts disease and provider from directory structure.
    
     Directory structure expected: data/new_data/PROVIDER/file.pdf
@@ -19,18 +19,18 @@ def load_pdf_documents(pdf_path: Path) -> List[Document]:
    
     Args:
         pdf_path: Path to the PDF file
+        api_key: LlamaCloud API key. If None, reads from LLAMA_CLOUD_API_KEY env variable
+               Get your API key from: https://cloud.llamaindex.ai/api-key
        
     Returns:
         List of Document objects with metadata (source, disease, provider, page_number)
     """
     try:
-       
         # Validate file exists
         if not pdf_path.exists():
             raise FileNotFoundError(f"PDF file not found at {pdf_path}")
        
         # Extract provider from directory structure
-        # Structure: data/new_data/PROVIDER/file.pdf
         path_parts = pdf_path.parts
         disease = "HBV"  # Default disease for this system
         provider = "unknown"
@@ -38,54 +38,225 @@ def load_pdf_documents(pdf_path: Path) -> List[Document]:
         # Find provider: it's the parent directory of the PDF file
         if len(path_parts) >= 2:
             provider = path_parts[-2]  # Parent directory (e.g., SASLT)
-            
+           
         # If provider is 'new_data', it means file is directly in new_data folder
         if provider.lower() == "new_data":
             provider = "unknown"
        
-        # Initialize PyMuPDF4LLMLoader
-        loader = PyMuPDF4LLMLoader(
-            str(pdf_path),
-            mode="page",
-            extract_images=True,
-            images_parser=TesseractBlobParser(),
-            table_strategy="lines"
+        # Get API key from parameter or environment variable
+        llama_api_key = api_key or os.getenv("LLAMA_CLOUD_API_KEY")
+        if not llama_api_key:
+            raise ValueError(
+                "LlamaCloud API key not found. Please provide api_key parameter or set "
+                "LLAMA_CLOUD_API_KEY environment variable. "
+                "Get your key from: https://cloud.llamaindex.ai/api-key"
+            )
+       
+        # Initialize LlamaParse with optimized settings for medical documents
+        parser = LlamaParse(
+            api_key=llama_api_key,
+            result_type="markdown",  # or "text" for plain text
+            verbose=True,
+            language="en",
+            # Medical document optimizations
+            parsing_instruction=(
+                "This is a medical guideline document. "
+                "Pay special attention to tables (including borderless tables), "
+                "clinical recommendations, dosage information, and reference citations. "
+                "Preserve table structure and maintain hierarchical headings."
+            ),
+            # Advanced options for better table handling
+            invalidate_cache=False,  # Use cache for faster re-processing
+            do_not_cache=False,
+            fast_mode=False,  # Use deep parsing for better accuracy
+            # Split by page for proper page numbering
+            split_by_page=True,  # This is the key parameter!
         )
        
-        raw_documents = loader.load()
+        # Parse the PDF file
+        logger.info(f"Parsing PDF with LlamaParse: {pdf_path.name}")
+        
+        # Use SimpleDirectoryReader with LlamaParse
+        file_extractor = {".pdf": parser}
+        reader = SimpleDirectoryReader(
+            input_files=[str(pdf_path)],
+            file_extractor=file_extractor
+        )
+        
+        # Load documents - each page will be a separate document when split_by_page=True
+        llama_documents = reader.load_data()
        
+        # Convert to LangChain Document format
         documents = []
-        for idx, doc in enumerate(raw_documents):
-            if doc.page_content.strip():
-                # Extract actual page number from metadata, default to sequential numbering
-                # PyMuPDF4LLMLoader uses 0-indexed pages, so we add 1 for human-readable page numbers
-                actual_page = doc.metadata.get("page")
-                if actual_page is not None:
-                    # If page is 0-indexed, add 1 to make it 1-indexed
-                    page_num = actual_page + 1 if actual_page == idx else actual_page
-                else:
-                    # Fallback to 1-indexed sequential numbering
-                    page_num = idx + 1
-                
-                processed_doc = Document(
-                    page_content=doc.page_content,
-                    metadata={
-                        "source": pdf_path.name,
-                        "disease": disease,
-                        "provider": provider,
-                        "page_number": page_num
-                    }
-                )
-                documents.append(processed_doc)
-
+        
+        for doc_idx, llama_doc in enumerate(llama_documents):
+            # When split_by_page=True, each llama_doc represents one page
+            # Check if page number exists in metadata, otherwise use index
+            page_num = llama_doc.metadata.get('page_number', doc_idx + 1)
+            
+            processed_doc = Document(
+                page_content=llama_doc.text.strip(),
+                metadata={
+                    "source": pdf_path.name,
+                    "disease": disease,
+                    "provider": provider,
+                    "page_number": page_num,
+                    "document_index": doc_idx,
+                    # Preserve any additional metadata from LlamaParse
+                    **{k: v for k, v in llama_doc.metadata.items() 
+                       if k not in ['source', 'disease', 'provider', 'page_number', 'document_index']}
+                }
+            )
+            documents.append(processed_doc)
+       
         logger.info(f"Loaded {len(documents)} pages from PDF: {pdf_path.name} (Disease: {disease}, Provider: {provider})")
         return documents
        
     except Exception as e:
         logger.error(f"Error loading PDF documents from {pdf_path}: {str(e)}")
         raise
+
+
+def load_pdf_documents_advanced(
+    pdf_path: Path,
+    api_key: str = None,
+    premium_mode: bool = False
+) -> List[Document]:
+    """
+    Advanced version with premium features for complex medical documents.
+    
+    Args:
+        pdf_path: Path to the PDF file
+        api_key: LlamaCloud API key
+        premium_mode: Use premium GPT-4o mode for highest accuracy (costs more)
+    
+    Returns:
+        List of Document objects with enhanced metadata
+    """
+    try:
+        if not pdf_path.exists():
+            raise FileNotFoundError(f"PDF file not found at {pdf_path}")
+       
+        path_parts = pdf_path.parts
+        disease = "HBV"
+        provider = path_parts[-2] if len(path_parts) >= 2 else "unknown"
+        if provider.lower() == "new_data":
+            provider = "unknown"
+       
+        llama_api_key = api_key or os.getenv("LLAMA_CLOUD_API_KEY")
+        if not llama_api_key:
+            raise ValueError("LlamaCloud API key required")
+       
+        # Advanced parser configuration
+        parser = LlamaParse(
+            api_key=llama_api_key,
+            result_type="markdown",
+            verbose=True,
+            language="en",
+            # Premium mode uses GPT-4o for better accuracy
+            premium_mode=premium_mode,
+            # Detailed parsing instructions for medical content
+            parsing_instruction=(
+                "Medical guideline document with complex tables. Instructions:\n"
+                "0. Keep the original text intact without changing anything\n"
+                "1. Preserve all table structures, especially borderless tables\n"
+                "2. Maintain hierarchical organization of sections and subsections\n"
+                "3. Keep dosage tables and treatment algorithms intact\n"
+                "4. Preserve reference numbers and citations\n"
+                "5. Identify and mark clinical recommendation levels\n"
+                "6. Extract figures and their captions accurately"
+            ),
+            # Extract structured data
+            take_screenshot=True,  # Capture page screenshots for reference
+            # Table-specific optimizations
+            invalidate_cache=False,
+            do_not_cache=False,
+            fast_mode=False,
+            # Critical: split by page for accurate page numbering
+            split_by_page=True,
+        )
+       
+        file_extractor = {".pdf": parser}
+        reader = SimpleDirectoryReader(
+            input_files=[str(pdf_path)],
+            file_extractor=file_extractor
+        )
+       
+        logger.info(f"Parsing PDF with LlamaParse (Premium: {premium_mode}): {pdf_path.name}")
+        llama_documents = reader.load_data()
+       
+        documents = []
+        for doc_idx, llama_doc in enumerate(llama_documents):
+            # Get page number from metadata or use index
+            page_num = llama_doc.metadata.get('page_number', doc_idx + 1)
+            
+            # Enhanced metadata
+            metadata = {
+                "source": pdf_path.name,
+                "disease": disease,
+                "provider": provider,
+                "page_number": page_num,
+                "document_index": doc_idx + 1,
+                "parser": "llamaparse",
+                "premium_mode": premium_mode
+            }
+           
+            # Add additional metadata from LlamaIndex document
+            if hasattr(llama_doc, 'metadata'):
+                # Merge additional metadata, avoiding duplicates
+                for key, value in llama_doc.metadata.items():
+                    if key not in metadata:
+                        metadata[key] = value
+           
+            processed_doc = Document(
+                page_content=llama_doc.text.strip(),
+                metadata=metadata
+            )
+            documents.append(processed_doc)
+       
+        logger.info(f"Loaded {len(documents)} pages from PDF: {pdf_path.name}")
+        return documents
+       
+    except Exception as e:
+        logger.error(f"Error loading PDF documents from {pdf_path}: {str(e)}")
+        raise
+
+
+# Batch processing function for multiple PDFs
+def load_multiple_pdfs(
+    pdf_directory: Path,
+    api_key: str = None,
+    file_pattern: str = "*.pdf"
+) -> List[Document]:
+    """
+    Load multiple PDF files from a directory.
     
+    Args:
+        pdf_directory: Directory containing PDF files
+        api_key: LlamaCloud API key
+        file_pattern: Glob pattern for PDF files (default: "*.pdf")
     
+    Returns:
+        List of all documents from all PDFs
+    """
+    all_documents = []
+    pdf_files = list(pdf_directory.glob(file_pattern))
+   
+    logger.info(f"Found {len(pdf_files)} PDF files to process")
+   
+    for pdf_path in pdf_files:
+        try:
+            documents = load_pdf_documents(pdf_path, api_key=api_key)
+            all_documents.extend(documents)
+            logger.info(f"Successfully processed: {pdf_path.name}")
+        except Exception as e:
+            logger.error(f"Failed to process {pdf_path.name}: {str(e)}")
+            continue
+   
+    logger.info(f"Total documents loaded: {len(all_documents)}")
+    return all_documents
+
+
 def load_markdown_documents(md_path: Path) -> List[Document]:
     """
     Load and process Markdown medical guidelines.
@@ -139,4 +310,5 @@ def load_markdown_documents(md_path: Path) -> List[Document]:
 
     except Exception as e:
         logger.error(f"Error loading Markdown document from {md_path}: {str(e)}")
-        raise
+        raise
+

core/hbv_assessment.py

@@ -113,110 +113,86 @@ def normalize_recommendations(text: str) -> str:
 
 # SASLT 2021 Guidelines - Hardcoded Page Contents
 SASLT_GUIDELINES = """
-===== TREATMENT RECOMMENDATIONS =====
-
+===== SASLT 2021 GUIDELINES: TREATMENT & MANAGEMENT =====
 
 ### 1. INITIATION OF TREATMENT [SASLT 2021, p. 6]
 
-• Treatment indications should also take into account patient's age, health status, risk of HBV transmission, family history of HCC or cirrhosis and extrahepatic manifestations
-
-• All patients with chronic hepatitis B (HBV DNA > 2,000 IU/mL, ALT > ULN), regardless of HBeAg status, and/or at least moderate liver necroinflammation or fibrosis (Grade A)
-
-• Patients with cirrhosis (compensated or decompensated), with any detectable HBV DNA level and regardless of ALT levels (Grade A)
-
-• HBV Eligibility Criteria: 
-    •Patients with HBV DNA > 20,000 IU/mL and ALT > 2xULN, regardless of the degree of fibrosis (Grade B)
-
-    • Patients with HBeAg-positive chronic HBV infection (persistently normal ALT and high HBV DNA levels) may be treated if they are > 30 years, regardless of the severity of liver histological lesions (Grade D)
-
-    • Patients with chronic HBV infection (HBV DNA > 2,000 IU/mL, ALT > ULN), regardless of HBeAg status, and a family history of HCC or cirrhosis and extrahepatic manifestations (Grade D)
+• Treatment indications should also take into account patient's age, health status, risk of HBV transmission, family history of HCC or cirrhosis and extrahepatic manifestations [SASLT 2021, p. 6]
+• All patients with chronic hepatitis B (HBV DNA > 2,000 IU/mL, ALT > ULN), regardless of HBeAg status, and/or at least moderate liver necroinflammation or fibrosis (Grade A) [SASLT 2021, p. 6]
+• Patients with cirrhosis (compensated or decompensated), with any detectable HBV DNA level and regardless of ALT levels (Grade A) [SASLT 2021, p. 6]
+• Patients with HBV DNA > 20,000 IU/mL and ALT > 2xULN, regardless of the degree of fibrosis (Grade B) [SASLT 2021, p. 6]
+• Patients with HBeAg-positive chronic HBV infection (persistently normal ALT and high HBV DNA levels) may be treated if they are > 30 years, regardless of the severity of liver histological lesions (Grade D) [SASLT 2021, p. 6]
+• Patients with chronic HBV infection (HBV DNA > 2,000 IU/mL, ALT > ULN), regardless of HBeAg status, and a family history of HCC or cirrhosis and extrahepatic manifestations (Grade D) [SASLT 2021, p. 6]
 
 
 ### 2. MANAGEMENT ALGORITHM [SASLT 2021, p. 6]
 
-• HBsAg positive with chronic HBV infection and no signs of chronic hepatitis → Monitor (HBsAg, HBeAg, HBV DNA, ALT, fibrosis assessment). Consider: risk of HCC, risk of HBV reactivation, extrahepatic manifestations, risk of HBV transmission
-
-• CHB (with/without cirrhosis) → Start antiviral treatment if indicated, otherwise return to monitoring
-
-• HBsAg negative, anti-HBc positive → No specialist follow-up (inform about HBV reactivation risk). In case of immunosuppression: start oral antiviral prophylaxis or monitor
+• HBsAg positive with chronic HBV infection and no signs of chronic hepatitis → Monitor (HBsAg, HBeAg, HBV DNA, ALT, fibrosis assessment). Consider: risk of HCC, risk of HBV reactivation, extrahepatic manifestations, risk of HBV transmission [SASLT 2021, p. 6]
+• CHB (with/without cirrhosis) → Start antiviral treatment if indicated, otherwise return to monitoring [SASLT 2021, p. 6]
+• HBsAg negative, anti-HBc positive → No specialist follow-up (inform about HBV reactivation risk). In case of immunosuppression: start oral antiviral prophylaxis or monitor [SASLT 2021, p. 6]
 
 
 ### 3. MONITORING OF UNTREATED PATIENTS [SASLT 2021, p. 6-7]
 
-• Patients with HBeAg-positive chronic HBV infection who are younger than 30 years should be followed at least every 3-6 months (Grade B)
-
-• Patients with HBeAg-negative chronic HBV infection and serum HBV DNA <2,000 IU/ml should be followed every 6-12 months (Grade B)
-
-• Patients with HBeAg-negative chronic HBV infection and serum HBV DNA ≥2,000 IU/ml should be followed every 3 months for the first year and thereafter every 6 months (Grade D)
+• Patients with HBeAg-positive chronic HBV infection who are younger than 30 years should be followed at least every 3-6 months (Grade B) [SASLT 2021, p. 7]
+• Patients with HBeAg-negative chronic HBV infection and serum HBV DNA <2,000 IU/ml should be followed every 6-12 months (Grade B) [SASLT 2021, p. 7]
+• Patients with HBeAg-negative chronic HBV infection and serum HBV DNA ≥2,000 IU/ml should be followed every 3 months for the first year and thereafter every 6 months (Grade D) [SASLT 2021, p. 7]
 
 
 ### 4. CHRONIC HEPATITIS B (CHB) TREATMENT [SASLT 2021, p. 7-8]
 
-• The treatment of choice is the long-term administration of a potent nucleos(t)ide analogue NA with a high barrier to resistance, regardless of the severity of liver disease (Grade A)
-
-• Preferred regimens are ETV, TDF and TAF as monotherapies (Grade A)
-
-• LAM, ADV and TBV are not recommended in the treatment of CHB (Grade A)
+• The treatment of choice is the long-term administration of a potent nucleos(t)ide analogue NA with a high barrier to resistance, regardless of the severity of liver disease (Grade A) [SASLT 2021, p. 8]
+• Preferred regimens are ETV, TDF and TAF as monotherapies (Grade A) [SASLT 2021, p. 8]
+• LAM, ADV and TBV are not recommended in the treatment of CHB (Grade A) [SASLT 2021, p. 8]
 
 
 ### 5. HBV-HCV COINFECTION [SASLT 2021, p. 8-9]
 
-• Treatment of HCV through DAAs may lead to reactivation of HBV. Patients who meet the criteria for HBV treatment should be treated concurrently or before initiation of DAA (Grade A)
-
-• HBV DNA and ALT should be monitored every four to eight weeks while on DAA and three months after completion of therapy (Grade D)
-
-• ALT level should be monitored every four weeks while on DAA for patients who are HBsAg-negative but HBcAb-positive. If ALT starts to rise, HBsAg and HBV DNA must be obtained to determine the need to start HBV treatment (Grade D)
+• Treatment of HCV through DAAs may lead to reactivation of HBV. Patients who meet the criteria for HBV treatment should be treated concurrently or before initiation of DAA (Grade A) [SASLT 2021, p. 9]
+• HBV DNA and ALT should be monitored every four to eight weeks while on DAA and three months after completion of therapy (Grade D) [SASLT 2021, p. 9]
+• ALT level should be monitored every four weeks while on DAA for patients who are HBsAg-negative but HBcAb-positive. If ALT starts to rise, HBsAg and HBV DNA must be obtained to determine the need to start HBV treatment (Grade D) [SASLT 2021, p. 9]
 
 
 ### 6. HBV-HDV COINFECTION [SASLT 2021, p. 9]
 
-• HDV is a defective virus that requires HBsAg to envelop its delta antigen, causing coinfection with HBV or superinfection in chronic HBV patients
-
-• Active HDV infection is defined by HDV IgM and RNA presence with unexplained LFT elevation
-
-• Treatment goal: Suppression of HDV replication
-
-• PEG-IFN for 1 year shows long-term benefits despite post-treatment viral relapse
-
-• NA monotherapy is ineffective against HDV replication
+• HDV is a defective virus that requires HBsAg to envelop its delta antigen, causing coinfection with HBV or superinfection in chronic HBV patients [SASLT 2021, p. 9]
+• Active HDV infection is defined by HDV IgM and RNA presence with unexplained LFT elevation [SASLT 2021, p. 9]
+• Treatment goal: Suppression of HDV replication [SASLT 2021, p. 9]
+• PEG-IFN for 1 year shows long-term benefits despite post-treatment viral relapse [SASLT 2021, p. 9]
+• NA monotherapy is ineffective against HDV replication [SASLT 2021, p. 9]
 
 
 ### 7. HBV-HIV COINFECTION [SASLT 2021, p. 9]
 
-• All HIV-positive patients with HBV co-infection should start ART irrespective of CD4 cell count (Grade A)
-
-• HBV-HIV co-infected patients should be treated with TDF- or TAF-based ART regimen (Grade A)
+• All HIV-positive patients with HBV co-infection should start ART irrespective of CD4 cell count (Grade A) [SASLT 2021, p. 9]
+• HBV-HIV co-infected patients should be treated with TDF- or TAF-based ART regimen (Grade A) [SASLT 2021, p. 9]
 
 
 ### 8. IMMUNOCOMPROMISED PATIENTS [SASLT 2021, p. 9]
 
-• Prophylaxis for all HBsAg-positive patients before chemotherapy or immunosuppressive therapy (Grade A)
-
-• HBsAg-negative/anti-HBc-positive patients need HBV prophylaxis if receiving anti-CD20 or stem cell transplantation
-
-• Continue prophylaxis for ≥6 months after immunosuppression (12 months for anti-CD20)
+• Prophylaxis for all HBsAg-positive patients before chemotherapy or immunosuppressive therapy (Grade A) [SASLT 2021, p. 9]
+• HBsAg-negative/anti-HBc-positive patients need HBV prophylaxis if receiving anti-CD20 or stem cell transplantation [SASLT 2021, p. 9]
+• Continue prophylaxis for ≥6 months after immunosuppression (12 months for anti-CD20) [SASLT 2021, p. 9]
+• All patients undergoing immunosuppressive treatment or chemotherapy, even short‑term courses, should be screened for HBsAg, anti‑HBc, and anti‑HBs (and HBV DNA, if HBsAg is already positive). [SASLT 2021, p. 9]
+• We recommend prophylaxis for all patients with positive HBsAg before initiating chemotherapy or other immunosuppressive agents. [SASLT 2021, p. 9]
+• For HBsAg-negative and anti-HBc positive patients, we recommend HBV prophylaxis if they are candidates for anti CD20 or are undergoing stem cell transplantation. [SASLT 2021, p. 9]
+• We recommend starting HBV prophylaxis for HBsAg or anti‑HBc positive patients undergoing treatment with tumor necrosis factor (TNF) inhibitors. [SASLT 2021, p. 9]
+• We recommend HBV prophylaxis for all patients who are HBsAg or anti-HBc positive before initiation of immunotherapy such as anti‑programmed cell death (PD) ‑1 and anti‑programmed cell death‑ligand 1 (PD‑L1) therapy. [SASLT 2021, p. 9]
 
 
 ### 9. PREGNANCY [SASLT 2021, p. 9-10]
 
-• Screen all pregnant women for HBV in first trimester (Grade A)
-
-• HBV vaccine is safe in pregnancy for non-immune women without chronic HBV
-
-• Treat pregnant women meeting standard therapy indications
-
-• Start antiviral prophylaxis with TDF (or TAF) for HBV DNA >100,000 IU/mL at 24-28 weeks (Grade D)
-
-• Switch to TDF/TAF if on ETV, ADV, or interferon during pregnancy (Grade D)
-
-• Delivery mode based on obstetric indications only
-
-• Breastfeeding permitted for HBsAg+ women on TDF (Grade B)
-
-----
+• Screen all pregnant women for HBV in first trimester (Grade A) [SASLT 2021, p. 9]
+• HBV vaccine is safe in pregnancy for non-immune women without chronic HBV. [SASLT 2021, p. 9]
+• Treat pregnant women meeting standard therapy indications [SASLT 2021, p. 9]
+• Start antiviral prophylaxis with TDF (or TAF) for HBV DNA >100,000 IU/mL at 24-28 weeks (Grade D) [SASLT 2021, p. 10]
+• Switch to TDF/TAF if on ETV, ADV, or interferon during pregnancy (Grade D) [SASLT 2021, p. 10]
+• Delivery mode based on obstetric indications only [SASLT 2021, p. 10]
+• Breastfeeding permitted for HBsAg+ women on TDF (Grade B) [SASLT 2021, p. 10]
 """
 
 
+
 def assess_hbv_eligibility(patient_data: Dict[str, Any]) -> Dict[str, Any]:
     """
     Assess patient eligibility for HBV treatment based on SASLT 2021 guidelines
@@ -231,12 +207,12 @@ def assess_hbv_eligibility(patient_data: Dict[str, Any]) -> Dict[str, Any]:
         - recommendations: str (comprehensive narrative with inline citations in format [SASLT 2021, Page X])
     """
     try:
-        # Check if HBsAg is positive (required for treatment consideration)
-        if patient_data.get("hbsag_status") != "Positive":
-            return {
-                "eligible": False,
-                "recommendations": "Patient is HBsAg negative. HBV treatment is not indicated. HBsAg positivity is required for HBV treatment consideration according to SASLT 2021 guidelines."
-            }
+        # # Check if HBsAg is positive (required for treatment consideration)
+        # if patient_data.get("hbsag_status") != "Positive":
+        #     return {
+        #         "eligible": False,
+        #         "recommendations": "Patient is HBsAg negative. HBV treatment is not indicated. HBsAg positivity is required for HBV treatment consideration according to SASLT 2021 guidelines."
+        #     }
         
         # Use hardcoded SASLT 2021 guidelines instead of RAG retrieval
         logger.info("Using hardcoded SASLT 2021 guidelines (Pages 3, 4, 6, 7, 8, 9, 10)")
@@ -262,7 +238,7 @@ def assess_hbv_eligibility(patient_data: Dict[str, Any]) -> Dict[str, Any]:
         comorbidities = patient_data.get("other_comorbidities", [])
         
         # Create prompt for LLM to analyze patient against guidelines
-        analysis_prompt = f"""You are an HBV treatment eligibility assessment system. Analyze the patient data against SASLT 2021 guidelines.
+        analysis_prompt = f"""You are an HBV treatment eligibility assessment system. Analyze the patient data against the SASLT 2021 guidelines.
 PATIENT DATA:
 - Sex: {sex}
 - Age: {age} years
@@ -290,19 +266,30 @@ You MUST respond with a valid JSON object in this exact format:
   "recommendations": "Comprehensive assessment with inline citations"
 }}
 IMPORTANT JSON FORMATTING:
-- Return ONLY valid JSON without markdown code blocks
+- Return ONLY valid JSON without markdown code blocks.
 - You MUST use "\\n" to indicate line breaks inside the "recommendations" string and format the content as clear bullet lists prefixed with "- ".
 - Do NOT include literal newline characters. Use \\n for every new bullet or line.
 - Use SINGLE \\n between lines. Do NOT use \\n\\n (double newlines) anywhere.
 
+CRITICAL ELIGIBILITY HIERARCHY:
+1.  **Check Special Populations FIRST (Pages 9-10):** "Eligible" means eligible for ANY antiviral intervention (treatment OR prophylaxis).
+    * If patient is **HBsAg-positive** AND **Immunosuppressed** (Page 9): Set **"eligible": true"** (for prophylaxis).
+    * If patient has **HBV-HIV coinfection** (Page 9): Set **"eligible": true"** (for ART).
+    * If patient is **Pregnant** with **HBV DNA > 100,000 IU/mL** (Page 10): Set **"eligible": true"** (for prophylaxis).
+    * If patient has **HBV-HCV coinfection** AND meets HBV treatment criteria (Page 9): Set **"eligible": true"**.
+2.  **Check Standard Criteria SECOND (Page 6):**
+    * If *not* eligible based on special populations, check standard criteria (HBV DNA, ALT, fibrosis, age, family history).
+    * If any Page 6 criteria are met (e.g., Cirrhosis, or HBV DNA > 2,000 + ALT > ULN + moderate fibrosis, etc.): Set **"eligible": true"**.
+3.  **If NO criteria from (1) or (2) are met:** Set **"eligible": false"**.
+4.  The "eligible" flag MUST be consistent with the "Eligibility and Rationale" bullet. Do not contradict yourself.
+
 STRUCTURE AND CONTENT OF "recommendations" (CONCISE & ORGANIZED):
 - Use ONLY these sections in this exact order, each as a header followed by 1-3 concise bullets:
-  1. "Eligibility and Rationale:" (1-2 bullets max)
-  2. "Treatment Recommendations:" (1-3 bullets: first-line drugs if eligible, or "Not applicable" if not eligible)
+  1. "Eligibility and Rationale:" (1-2 bullets max, MUST state the primary reason for eligibility/ineligibility)
+  2. "Treatment Recommendations:" (1-3 bullets: first-line drugs if eligible, or "Treatment not indicated" if not eligible)
   3. "Monitoring and Follow-up:" (1-2 bullets)
-  4. "Special Considerations:" (0-2 bullets, ONLY if applicable to this patient)
+  4. "Special Considerations:" (0-2 bullets, ONLY if applicable and *not* the primary reason for eligibility)
   5. "References:" (1 line listing pages cited)
-
 - OMIT "Additional Notes" and any other sections.
 - Keep each bullet to ONE sentence (max 25 words per bullet).
 - Total output: aim for 8-12 bullets maximum across all sections.
@@ -313,31 +300,49 @@ BULLETING AND CITATIONS RULES:
 - Only cite pages 6–10 that actually contain the information.
 
 STRICT ACCURACY AND CONSISTENCY RULES:
-- Do NOT contradict yourself. Ensure eligibility conclusion matches all bullets.
-- Use ONLY the provided SASLT 2021 content; do NOT add external knowledge.
-- Numeric thresholds must match the guideline text exactly.
-- If info is not in the provided pages, do not state it or cite it.
-- BREVITY: Each bullet = 1 sentence, max 25 words. Total = 8-12 bullets max.
+- **NO CONTRADICTIONS:** The "eligible" flag MUST match the rationale. Follow the CRITICAL ELIGIBILITY HIERARCHY.
+- **Rationale First:** The Rationale MUST state the primary reason. If eligible due to immunosuppression, state that, even if Page 6 criteria are not met.
+- **Use ONLY the provided SASLT 2021 content;** do NOT add external knowledge.
+- **BREVITY:** Each bullet = 1 sentence, max 25 words. Total = 8-12 bullets max.
 
 PAGE-TO-TOPIC MAPPING GUIDANCE (for correct citations):
-- Page 6: Initiation of treatment, management algorithm, start of monitoring of untreated patients.
-- Page 7: Monitoring of untreated patients (continuation), CHB treatment principles.
+- Page 6: Standard initiation of treatment criteria, management algorithm.
+- Page 7: Monitoring of untreated patients, CHB treatment principles.
 - Page 8: Treatment drugs/regimens (ETV, TDF, TAF), agents not recommended.
-- Page 9: Special populations (HBV-HCV, HBV-HDV, HBV-HIV, immunocompromised).
-- Page 10: Pregnancy-related recommendations (late pregnancy prophylaxis, breastfeeding, switching).
-
-EXAMPLE OUTPUT (ELIGIBLE PATIENT) - Use SINGLE \\n only:
-Eligibility and Rationale:\n- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, moderate fibrosis (Grade A) [SASLT 2021, Page 6]\nTreatment Recommendations:\n- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]\nMonitoring and Follow-up:\n- Monitor treatment response per SASLT protocol [SASLT 2021, Page 7]\nSpecial Considerations:\n- HBV-HIV coinfection: use TDF- or TAF-based ART (Grade A) [SASLT 2021, Page 9]\nReferences:\n- Pages 6, 7, 8, 9: Treatment criteria, drugs, monitoring, HIV coinfection
-
-EXAMPLE OUTPUT (NOT ELIGIBLE PATIENT) - Use SINGLE \\n only:
-Eligibility and Rationale:\n- Not eligible: HBV DNA < 2,000 IU/mL, ALT ≤ ULN, no significant fibrosis [SASLT 2021, Page 6]\nTreatment Recommendations:\n- Treatment not indicated at this time [SASLT 2021, Page 6]\nMonitoring and Follow-up:\n- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7]\nReferences:\n- Pages 6, 7: Treatment criteria, monitoring protocols
+- Page 9: Special populations (HBV-HCV, HBV-HDV, HBV-HIV, Immunocompromised).
+- Page 10: Pregnancy-related recommendations.
 
+---
+EXAMPLE OUTPUT (ELIGIBLE - STANDARD CRITERIA) - Use SINGLE \\n only:
+{{
+  "eligible": true,
+  "recommendations": "Eligibility and Rationale:\\n- Eligible: HBV DNA > 2,000 IU/mL, ALT > ULN, moderate fibrosis (Grade A) [SASLT 2021, Page 6]\\nTreatment Recommendations:\\n- Start monotherapy with ETV, TDF, or TAF (Grade A) [SASLT 2021, Page 8]\\nMonitoring and Follow-up:\\n- Monitor treatment response per SASLT protocol [SASLT 2021, Page 7]\\nReferences:\\n- Pages 6, 7, 8: Treatment criteria, drugs, monitoring"
+}}
+---
+EXAMPLE OUTPUT (NOT ELIGIBLE - STANDARD CRITERIA) - Use SINGLE \\n only:
+{{
+  "eligible": false,
+  "recommendations": "Eligibility and Rationale:\\n- Not eligible: HBV DNA < 2,000 IU/mL, ALT ≤ ULN, no significant fibrosis [SASLT 2021, Page 6]\\nTreatment Recommendations:\\n- Treatment not indicated at this time [SASLT 2021, Page 6]\\nMonitoring and Follow-up:\\n- Monitor every 6-12 months (HBeAg-negative, HBV DNA < 2,000 IU/mL) (Grade B) [SASLT 2021, Page 7]\\nReferences:\\n- Pages 6, 7: Treatment criteria, monitoring protocols"
+}}
+---
+EXAMPLE OUTPUT (ELIGIBLE - IMMUNOSUPPRESSION) - Use SINGLE \\n only:
+{{
+  "eligible": true,
+  "recommendations": "Eligibility and Rationale:\\n- Eligible: HBsAg-positive patient requires prophylaxis for immunosuppressive therapy (Grade A) [SASLT 2021, Page 9]\\nTreatment Recommendations:\\n- Start antiviral prophylaxis (e.g., ETV, TDF, TAF) [SASLT 2021, Page 8, 9]\\nMonitoring and Follow-up:\\n- Continue prophylaxis for ≥6 months after immunosuppression (12 for anti-CD20) [SASLT 2021, Page 9]\\nReferences:\\n- Pages 8, 9: Prophylaxis criteria, drug options, monitoring"
+}}
+---
+EXAMPLE OUTPUT (ELIGIBLE - HIV COINFECTION) - Use SINGLE \\n only:
+{{
+  "eligible": true,
+  "recommendations": "Eligibility and Rationale:\\n- Eligible: Patient has HBV-HIV coinfection and should start ART (Grade A) [SASLT 2021, Page 9]\\nTreatment Recommendations:\\n- ART regimen must include TDF- or TAF-based therapy (Grade A) [SASLT 2021, Page 9]\\nMonitoring and Follow-up:\\n- Monitor patient closely after ART initiation for immune reconstitution [SASLT 2021, Page 9]\\nReferences:\\n- Page 9: HBV-HIV coinfection management, ART regimens"
+}}
+---
 CRITICAL REQUIREMENTS:
 1. Base assessment ONLY on SASLT 2021 guidelines provided.
-2. Keep output SHORT: 8-12 bullets total, 1 sentence per bullet (max 25 words).
-3. Use ONLY the 5 sections listed above (omit "Additional Notes" and "Criteria Met/Not Met").
-4. Cite exact page at end of each bullet: [SASLT 2021, Page X].
-5. End with "References:" listing pages in ascending order.
+2. Follow the CRITICAL ELIGIBILITY HIERARCHY to avoid contradictions.
+3. Keep output SHORT: 8-12 bullets total, 1 sentence per bullet (max 25 words).
+4. Use ONLY the 5 sections listed above.
+5. Cite exact page at end of each bullet: [SASLT 2021, Page X].
 6. Return ONLY valid JSON, no markdown, no extra text.
 """
         

core/utils.py

@@ -9,7 +9,7 @@ from typing import List, Optional, Iterable
 from langchain.schema import Document
 from langchain_community.vectorstores import FAISS
 
-from .config import get_embedding_model, VECTOR_STORE_DIR, CHUNKS_PATH, NEW_DATA, PROCESSED_DATA
+from .config import get_embedding_model, VECTOR_STORE_DIR, CHUNKS_PATH, NEW_DATA, PROCESSED_DATA, settings
 from .text_processors import markdown_splitter, recursive_splitter
 from . import data_loaders
 
@@ -126,7 +126,15 @@ def _load_documents_for_file(file_path: Path) -> List[Document]:
     """
     try:
         if file_path.suffix.lower() == '.pdf':
-            return data_loaders.load_pdf_documents(file_path)
+            # Use advanced LlamaParse loader with settings from config
+            api_key = settings.LLAMA_CLOUD_API_KEY
+            premium_mode = settings.LLAMA_PREMIUM_MODE
+            
+            return data_loaders.load_pdf_documents_advanced(
+                file_path,
+                api_key=api_key,
+                premium_mode=premium_mode
+            )
         return data_loaders.load_markdown_documents(file_path)
     except Exception as e:
         logger.error(f"Failed to load {file_path}: {e}")

data/HBV_Eligibility_TestCases - To Be Tested(Sheet1).csv

@@ -0,0 +1,31 @@
+Case ID,Age,Sex,Pregnancy Status,HBsAg,HBeAg,HBV DNA (IU/mL),ALT (U/L),Fibrosis/Cirrhosis Stage,Necroinflammation,Extrahepatic Manifestations,Immunosuppressive Therapy,Coinfections,Family History of HCC/Cirrhosis,Smoking,Comorbidities,Eligibility,Rationale
+Case1,28,M,No,Positive (36 months),Positive,8500000,120,F1,Moderate,No,None,None,No,No,None,,
+Case2,45,F,No,Positive (60 months),Negative,15000,65,F1�F2,Moderate,No,None,None,No,No,Metabolic syndrome,,
+Case3,52,M,No,Positive (120 months),Negative,3500,32,F2�F3,Mild,No,None,None,No,No,None,,
+Case4,60,M,No,Positive (240 months),Negative,700,28,F4 (compensated),Mild,No,None,None,No,No,None,,
+Case5,22,F,No,Positive (18 months),Positive,25000000,27,F0,None,No,None,None,No,No,None,,
+Case6,40,M,No,Positive (84 months),Negative,900,22,F0�F1,None,No,None,None,No,No,None,,
+Case7,31,F,Yes (28 weeks),Positive (48 months),Positive,500000,35,F1,Mild,No,None,None,No,No,None,,
+Case8,35,M,No,Positive (72 months),Positive,9000000,33,F1,Mild,No,None,None,No,No,None,,
+Case9,55,F,No,Positive (180 months),Negative,6200,45,F1�F2,Moderate,No,None,None,Yes (father HCC),No,None,,
+Case10,63,M,No,Positive (300 months),Negative,50,32,F4 (decompensated),Mild,No,None,None,No,No,None,,
+Case11,68,M,No,Positive (25 years),Negative,5800,41,F2,Mild,No,None,None,No,No,Diabetes,,
+Case12,50,F,No,Positive,Negative,450,32,N/A,Mild,No,Yes (tacrolimus),None,No,No,Post-transplant,,
+Case13,27,M,No,Positive (36 months),Negative,1700,55,F1,Mild,No,None,None,No,No,None,,
+Case14,44,F,No,Positive,Negative,3000,70,F2,Moderate,No,None,HCV RNA positive,No,No,None,,
+Case15,33,M,No,Positive,Negative,25000,30,F1,Mild,No,None,HIV positive,No,No,None,,
+Case16,30,F,Yes (30 weeks),Positive,Positive,40000,28,F0,None,No,None,None,No,No,None,,
+Case17,39,M,No,Positive,Negative,4800,250,F1,Marked,No,None,None,No,No,None,,
+Case18,49,F,No,Positive,Negative,2300,37,F1,Mild,No,None,None,Yes (mother cirrhosis),No,None,,
+Case19,56,M,No,Positive,Negative,10,30,F4 (compensated),Mild,No,None,None,No,No,None,,
+Case20,41,F,No,Positive (10 years),Negative,1900,28,F1,Mild,No,None,None,No,No,None,,
+Case21,56,F,No,Positive,Negative,0,20,F4 (compensated),Mild,No,None,None,No,No,None,,
+Case22,70,M,No,Positive (30 years),Negative,1800,27,F2,Mild,No,None,None,No,Yes (smoker),Hypertension,,
+Case23,33,F,Yes (12 weeks),Positive,Positive,300000,40,F1,Mild,No,None,None,No,No,None,,
+Case24,46,M,No,Positive (8 years),Negative,50000,48,F2,Moderate,Yes (vasculitis),None,None,No,No,None,,
+Case25,58,F,No,Positive,Negative,2000,60,F2,Moderate,No,None,None,No,No,CKD stage 3,,
+Case26,29,M,No,Positive,Positive,2500000,33,F0,None,No,None,None,No,No,None,,
+Case27,54,M,No,Positive,Negative,750,18,F4 (compensated),Mild,No,None,None,No,No,None,,
+Case28,38,F,No,Positive,Negative,6000,80,F2,Moderate,No,None,None,No,No,Obesity,,
+Case29,42,M,No,Positive,Negative,1200,25,F1,Mild,No,None,None,No,No,None,,
+Case30,25,F,No,Positive (2 years),Positive,12000000,95,F1,Moderate,No,None,None,No,No,None,,

test_assessment_fixed.py

@@ -0,0 +1,238 @@
+import pandas as pd
+import requests
+import json
+import re
+from typing import Optional, List
+
+
+def parse_hbsag_duration(hbsag_value: str) -> int:
+    """Extract duration in months from HBsAg status string."""
+    if pd.isna(hbsag_value):
+        return 6  # Default to 6 months if not specified
+    
+    # Extract number and unit from strings like "Positive (36 months)" or "Positive (10 years)"
+    match = re.search(r'\((\d+)\s*(months?|years?)\)', str(hbsag_value))
+    if match:
+        value = int(match.group(1))
+        unit = match.group(2).lower()
+        return value * 12 if 'year' in unit else value
+    
+    # If just "Positive" with no duration, default to 6 months
+    return 6
+
+
+def parse_status(value: str) -> str:
+    """Parse status values to 'Positive' or 'Negative' (exact capitalization required)."""
+    if pd.isna(value):
+        return "Negative"
+    val_lower = str(value).lower()
+    if 'positive' in val_lower:
+        return "Positive"
+    elif 'negative' in val_lower:
+        return "Negative"
+    return "Negative"
+
+
+def parse_sex(value: str) -> str:
+    """Parse sex to 'Male' or 'Female' (exact capitalization required)."""
+    if pd.isna(value):
+        return "Male"
+    val_lower = str(value).lower()
+    if val_lower in ['m', 'male']:
+        return "Male"
+    elif val_lower in ['f', 'female']:
+        return "Female"
+    return "Male"
+
+
+def parse_pregnancy_status(sex: str, value: str) -> str:
+    """Parse pregnancy status to 'Not pregnant' or 'Pregnant' (exact capitalization required)."""
+    if sex == "Male":
+        return "Not pregnant"
+    if pd.isna(value):
+        return "Not pregnant"
+    val_lower = str(value).lower()
+    if 'yes' in val_lower or 'pregnant' in val_lower:
+        return "Pregnant"
+    return "Not pregnant"
+
+
+def parse_boolean(value: str) -> bool:
+    """Parse Yes/No values to boolean."""
+    if pd.isna(value):
+        return False
+    val_lower = str(value).lower()
+    return 'yes' in val_lower
+
+
+def parse_fibrosis_stage(value: str) -> str:
+    """Extract fibrosis stage - must be 'F0-F1', 'F2-F3', or 'F4'."""
+    if pd.isna(value) or value == "N/A":
+        return "F0-F1"
+    
+    val_str = str(value).upper()
+    
+    # Map specific values
+    if 'F4' in val_str or 'CIRRHOSIS' in val_str.upper():
+        return "F4"
+    elif 'F3' in val_str or 'F2' in val_str:
+        return "F2-F3"
+    elif 'F1' in val_str or 'F0' in val_str:
+        return "F0-F1"
+    
+    return "F0-F1"
+
+
+def parse_necroinflammation(value: str) -> str:
+    """Parse necroinflammation activity - must be 'A0', 'A1', 'A2', or 'A3'."""
+    if pd.isna(value) or str(value).lower() == "none":
+        return "A0"
+    
+    val_str = str(value).upper()
+    
+    # Map specific values
+    if 'A3' in val_str or 'SEVERE' in val_str:
+        return "A3"
+    elif 'A2' in val_str or 'MODERATE' in val_str:
+        return "A2"
+    elif 'A1' in val_str or 'MILD' in val_str:
+        return "A1"
+    elif 'A0' in val_str or 'MINIMAL' in val_str:
+        return "A0"
+    
+    return "A1"
+
+
+def parse_immunosuppression(value: str) -> str:
+    """Parse immunosuppression therapy status - must be 'None', 'Chemotherapy', or 'Other'."""
+    if pd.isna(value) or str(value).lower() == "none":
+        return "None"
+    
+    val_lower = str(value).lower()
+    if 'chemo' in val_lower:
+        return "Chemotherapy"
+    elif 'none' in val_lower:
+        return "None"
+    else:
+        return "Other"
+
+
+def parse_coinfections(value: str) -> List[str]:
+    """Parse coinfections - must be from list: HIV, HCV, HDV."""
+    if pd.isna(value) or str(value).lower() == "none":
+        return []
+    
+    coinfections = []
+    val_upper = str(value).upper()
+    
+    if 'HCV' in val_upper:
+        coinfections.append("HCV")
+    if 'HIV' in val_upper:
+        coinfections.append("HIV")
+    if 'HDV' in val_upper:
+        coinfections.append("HDV")
+    
+    return coinfections
+
+
+def parse_comorbidities(value: str) -> Optional[List[str]]:
+    """Parse other comorbidities."""
+    if pd.isna(value) or str(value).lower() == "none":
+        return None
+    return [str(value)]
+
+
+def create_api_payload(row: pd.Series) -> dict:
+    """Create API request payload from CSV row."""
+    sex = parse_sex(row['Sex'])
+    
+    return {
+        "sex": sex,
+        "age": int(row['Age']),
+        "pregnancy_status": parse_pregnancy_status(sex, row['Pregnancy Status']),
+        "hbsag_status": parse_status(row['HBsAg']),
+        "duration_hbsag_months": parse_hbsag_duration(row['HBsAg']),
+        "hbv_dna_level": float(row['HBV DNA (IU/mL)']),
+        "hbeag_status": parse_status(row['HBeAg']),
+        "alt_level": float(row['ALT (U/L)']),
+        "fibrosis_stage": parse_fibrosis_stage(row['Fibrosis/Cirrhosis Stage']),
+        "necroinflammatory_activity": parse_necroinflammation(row['Necroinflammation']),
+        "extrahepatic_manifestations": parse_boolean(row['Extrahepatic Manifestations']),
+        "immunosuppression_status": parse_immunosuppression(row['Immunosuppressive Therapy']),
+        "coinfections": parse_coinfections(row['Coinfections']),
+        "family_history_cirrhosis_hcc": parse_boolean(row['Family History of HCC/Cirrhosis']),
+        "other_comorbidities": parse_comorbidities(row['Comorbidities'])
+    }
+
+
+def assess_case(payload: dict, api_url: str) -> dict:
+    """Call the API to assess eligibility."""
+    try:
+        response = requests.post(
+            api_url,
+            json=payload,
+            headers={'Content-Type': 'application/json'},
+            timeout=30
+        )
+        response.raise_for_status()
+        return response.json()
+    except requests.exceptions.HTTPError as e:
+        # Try to get detailed error message
+        try:
+            error_detail = response.json()
+            print(f"API Error Details: {json.dumps(error_detail, indent=2)}")
+        except:
+            print(f"API Error: {e}")
+        return {"eligible": None, "recommendations": f"Error: {str(e)}"}
+    except requests.exceptions.RequestException as e:
+        print(f"API Error: {e}")
+        return {"eligible": None, "recommendations": f"Error: {str(e)}"}
+
+
+def main():
+    # Configuration
+    input_file = r"D:\Work\HBV AI Assistant\data\HBV_Eligibility_TestCases - To Be Tested(Sheet1).csv"
+    output_file = "HBV_Eligibility_Results.csv"
+    api_url = "https://moazx-hbv-ai-assistant.hf.space/assess"
+    
+    # Read CSV
+    print(f"Reading {input_file}...")
+    df = pd.read_csv(input_file, encoding='windows-1252')
+    
+    # Add columns for results
+    df['Eligibility'] = None
+    df['Rationale'] = None
+    
+    # Process each case
+    print(f"\nProcessing {len(df)} cases...")
+    for idx, row in df.iterrows():
+        case_id = row['Case ID']
+        print(f"\nProcessing {case_id}...")
+        
+        # Create payload
+        payload = create_api_payload(row)
+        print(f"Payload: {json.dumps(payload, indent=2)}")
+        
+        # Call API
+        result = assess_case(payload, api_url)
+        print(f"Result: {result}")
+        
+        # Update dataframe
+        df.at[idx, 'Eligibility'] = result.get('eligible')
+        df.at[idx, 'Rationale'] = result.get('recommendations', '')
+    
+    # Save results
+    print(f"\nSaving results to {output_file}...")
+    df.to_csv(output_file, index=False)
+    print("Done!")
+    
+    # Print summary
+    eligible_count = df['Eligibility'].sum() if df['Eligibility'].notna().any() else 0
+    print(f"\nSummary:")
+    print(f"Total cases: {len(df)}")
+    print(f"Eligible: {eligible_count}")
+    print(f"Not eligible: {len(df) - eligible_count}")
+
+
+if __name__ == "__main__":
+    main()