Purpose The immunologic properties of tumors can change during the clinical course. We aimed to compare the expression of PD-L1 and the infiltration of CD8+ tumor-infiltrating lymphocytes (TILs) in the tumor microenvironment between initial and recurrent head and neck squamous-cell carcinomas (HNSCCs). We also evaluated whether the changes of those immunologic properties in recurrent HNSCCs affect the oncologic outcome. Methods We included 42 patients who had been treated for both initial and recurrent HNSCCs. Pathologic specimens from initial and salvage surgery were obtained for each patient, and IHC staining of CD3, CD8, PD-1, and PD-L1 were done. Also, neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were collected. The change of each immunologic profile was expressed as the recurrent-to-initial ratio (R/I ratio). Results The change of CD3+, CD8+, PD-1, and PD-L1 positive cells varied widely. For more than half the patients, those values decreased in recurrent tumors. The median R/I ratio was 0.5 for CD3+ TILs, 0.6 for CD8+ TILs, 0.5 for PD-1 and 0.4 for PD-L1. In contrast, NLR and PLR increased in recurrent tumors for more than half the patients. The median R/I ratio of NLR and PLR was 1.6 and 1.4, respectively. In multivariate analysis, increased CD8 (R/I ratio >1) was the independent prognostic factor for better OS (hazard ratio 0.228; 95% CI 0.067 – 0.777; p= 0.018). Conclusion The change of immunologic properties along with the recurrence of HNSCC varied widely from patient to patient. Generally, the intratumoral biomarkers decreased, while the systemic inflammatory markers increased. The increased CD8+ TILs in recurrent HNSCCs was the significant prognostic factor for better overall survival.

No abstract

Objective: Clostridium difficile infection (CDI) is affected by various factors. The association between CDI and antibiotic use is well known, and proton pump inhibitors (PPI) and histamine H2 receptor antagonist (H2RA) are also considered to be causative agents. The Korea Institute of Drug Safety and Risk Management has established the Medical record Observation and Assessment for drug safety Network (MOA-Net) using the common data model (CDM). We analyzed the risk factors of CDI in patients using multi-institutional hospital medical records obtained through the MOA-Net. Methods: We conducted a nested case-control study for CDI patients. Each case of CDI was matched with four controls without CDI and adjusted for sex, age, and the admission period. Patients were considered to have CDI if the disease code (A047) was present in their medical records, or if they had any clostridium difficile tests. Comorbidities, laboratory values, and the use and duration of PPI and H2RA were included in the analysis. Results: A total of 3,782 patients and 15,128 matched controls were enrolled in six participating hospitals. The occurrence of CDI was associated with the use of PPI (odds ratio [OR] 1.21, 95% confidence interval [CI] 1.02-1.43) and H2RA (OR 2.18, 95% CI 1.85-2.56). Conditional logistic regression analyses showed that longer antibiotic use, H2RAs, a history of renal disease, decreased levels of albumin and platelets, and leukocytosis were related to CDI occurrence. Conclusion: Our study identified risk factors for CDI in Korean inpatients. We showed that analysis using multi-institutional electronic medical records through the CDM and MOA-Net is feasible.