Study Paper Ideas for Your TB Dataset

Idea 1 - Diagnostic Concordance Study (Most Publishable)

• Use culture as the reference standard
• Calculate sensitivity, specificity, PPV, NPV for each test (AFB, GeneXpert, histopath) individually and in combination
• Identify cases where histopath shows granulomas but GeneXpert/culture are negative (culture-negative TB) - these are high-value outliers

Idea 2 - Added Value of Combined Testing

• Define diagnostic algorithms (histopath alone vs histopath + GeneXpert vs all four tests combined)
• Calculate area under ROC curve for each algorithm
• Identify the minimum combination that achieves best yield with least cost/time

Idea 3 - Histopathological Pattern Analysis Correlated with Microbiological Confirmation

• Classify histopath findings: caseating vs non-caseating granuloma, Langhans giant cells, necrosis grade
• Correlate each pattern with probability of GeneXpert/AFB/culture positivity
• Answer: which histopath patterns most reliably predict microbiological confirmation?

Idea 4 - Culture-Negative TB Analysis

• Isolate the subset where histopath says TB but culture is negative
• Within that group, check how often GeneXpert was positive (rescuing the diagnosis)
• Describe clinical profile of these patients

Idea 5 - Site-Specific Subgroup Analysis (if your data includes multiple tissue sites)

• Stratify by specimen site (lymph node, pleural biopsy, bone, intestinal, etc.)
• Compare GeneXpert sensitivity by site against culture as reference
• Report which sites show highest and lowest GeneXpert yield

My Top Recommendation

• You already have all four data points per patient - no additional data collection needed
• Diagnostic concordance with histopath pattern correlation is novel, clinically useful, and fits neatly into journals like Diagnostic Microbiology & Infectious Disease, PLOS ONE, Indian Journal of Pathology and Microbiology, or BMC Infectious Diseases
• Sample size of even 100-150 cases is sufficient for this design

Suggested Study Design (Quick Template)

Core Research Angle

Paper Idea 1 - The Main Study (Strongest)

• From your dataset, identify all cases where histopath showed granulomatous inflammation / caseating necrosis (suggestive of TB) but no microbiology was ordered (no AFB, no GeneXpert, no culture)
• Calculate: what % of TB-suspicious histopath cases had zero microbiological workup?
• Among those where micro WAS done, how often did it add (confirmed) or subtract (ruled out) TB?
• Back-calculate: if micro had been done in the "path only" group, how many additional confirmations or rule-outs might have resulted?

• Directly exposes a clinical practice gap - actionable for hospital policy
• No one has quantified this systematically in an Indian tertiary care setting
• A 2025 meta-analysis on female genital TB shows histopath alone is insufficient - your study provides the real-world "why isn't micro being done?" data

Paper Idea 2 - Narrow but High Impact

• Take all your "path only" TB-suspicious cases
• Track final clinical diagnosis and treatment decision in each case
• Was anti-TB treatment started empirically? Was it appropriate? Were any patients treated for TB who didn't have it (no confirmation ever obtained)?
• Describe outcomes: delayed diagnosis, empirical treatment, missed alternative diagnoses (sarcoidosis, fungal, foreign body granuloma)

• Granulomatous inflammation has a broad differential - histopath alone cannot distinguish TB from sarcoid, histoplasma, leprosy, Crohn's
• "Path only" cases are being managed blindly - quantifying this is a patient safety issue
• Strong narrative for a policy/quality improvement paper

Paper Idea 3 - Specialty-Specific Focus

• Break down your dataset by tissue site (lymph node, pleura, bone, intestine, peritoneum, genital tract, etc.)
• For each site: what % had path only vs path + micro?
• Which surgical specialties (surgery, gynecology, orthopedics, gastroenterology) were most likely to send path only and skip micro?
• Identify the "highest offenders" - sites and specialties where the gap is worst

• Site-specific and specialty-specific data is very useful for targeted education
• A 2025 study on cervical lymph node TB (PMID: 40995226) found that pathological subtype and sampling strategy directly affect diagnostic yield - your audit adds the "who ordered what" layer
• This is an audit/quality improvement paper - easier ethics approval, shorter turnaround

Paper Idea 4 - Process Improvement Angle

• This is a before-after study if your hospital has (or can implement) a protocol change
• Phase 1 (retrospective): quantify the gap using existing data
• Phase 2 (prospective): after implementing a standing order that any granulomatous tissue goes to both path and micro, measure how yield improves
• Compare TB confirmation rates before and after

• Interventional quality improvement studies are high-value and practically impactful
• Easily publishable in Quality in Primary Care, BMJ Open, IJMM

My Top Pick for Your Dataset

What Makes This Novel

• Very few papers have looked at this from the ordering clinician's practice gap angle
• Most TB diagnostic studies compare test performance - yours exposes the system failure before tests even get ordered
• In India/high-burden settings, this is a real day-to-day problem with direct policy implications
• Your data is already collected - this is a retrospective audit, which means low effort, high output