Search

Objectives To assess the utility and performance of the large language model ChatGPT 4.0 regarding accuracy, completeness, and its potential as a time-saving tool for pathologists and laboratory directors. Methods A deidentified database of questions previously sent to pathology residents from health care providers was used as a source of general knowledge–type pathology questions. These questions were submitted to the large language model and the responses graded by subject matter experts in different pathology subspecialties. The grading criteria assessed accuracy, completeness, and the potential time savings for helping the pathologist craft the response. Results Overall, respondents thought that most of the answers would take less than 5 minutes of additional work to be used (85%). Accuracy and completeness for the 61 questions was high, with 98% of responses being at least "completely or mostly accurate" and 82% of responses "containing all relevant information." Of the respondents, 97% stated that the response would have "zero or near-zero potential for medical harm," and all thought the tool had potential to save time in constructing answers to health care providers' queries. Performance was similar in both Anatomic Pathology (AP) and Clinical Pathology (CP), with the only exception being some relevant information was excluded in 46% of AP answers vs only 10% in CP (P <.01). Conclusions ChatGPT version 4.0 gave responses that were predominantly accurate and complete for general knowledge–type pathology questions. With further research and when reviewed by a pathologist or laboratorian, this could facilitate its use as a pathologist's aid in answering questions from health care providers. [ABSTRACT FROM AUTHOR]

Background/Aims The ability of the pretreatment lymphocyte to monocyte ratio (LMR) to predict outcomes of patients with hepatocellular carcinoma (HCC) receiving sorafenib is not conclusively determined. Methods We retrospectively studied patients treated with sorafenib for HCC in two tertiary referral centers in South Korea (KOR) and United States (US). Primary endpoints were overall survival (OS) and progression-free survival (PFS). Predictive factors for the primary outcomes were determined by Cox proportional hazards models. A risk-assessment tool was developed. Results Compared to the US cohort, the KOR cohort was more heavily pretreated (72.1% vs 35.2%, p<0.001), had higher proportion with hepatitis B virus infection (87.6% vs 5.6%, p<0.001), and more distant metastases (83.2% vs 25.4%, p<0.001). Lower monocyte count in the KOR cohort (median, 462.7 vs 600.0/μL; p=0.023) resulted in a higher LMR (median, 2.6 vs 1.8; p<0.001). High LMR was associated with a significantly higher OS (hazard ratio [HR], 0.90; 95% confidence interval [CI], 0.83 to 0.98; p=0.021). This was confirmed in a sensitivity analysis including only Asian patients (HR, 0.90; 95% CI, 0.82 to 0.98; p=0.019). A nomogram for OS was constructed with following variables selected in the multivariate Cox model: LMR, race, previous treatment, log10AFP, lymph node metastasis, and Child-Pugh score. The concordance score was 0.71 (95% CI, 0.67 to 0.75). LMR did not predict PFS. Conclusions Pretreatment LMR predicts OS in HCC patients treated with sorafenib. Our OS nomogram, incorporating LMR, can be offered to clinicians to improve their ability to assess prognosis, strengthen the prognosis-based decision making, and inform patients in the clinic. [ABSTRACT FROM AUTHOR]

Background: Current staging systems for gallbladder cancer (GBC) are primarily based on surgical pathology and therefore are not relevant for unresectable patients and those undergoing neoadjuvant chemotherapy., Methods: Patients with a confirmed diagnosis of GBC managed at a tertiary referral center (2000-2016) were included. Independent predictors of overall survival (OS) were identified using multivariable analysis (MVA). A combination of these variables was then assessed to identify a set of factors that provided maximal accuracy in predicting OS, and a nomogram and a new staging system were created based on these factors. Harrell's C-statistic was calculated to evaluate the predictive accuracy of the nomogram and staging system., Results: A total of 528 patients were included in the final analysis. On MVA, factors predictive of poor OS were older age, ECOG performance status, hemoglobin level <9 g/dL, presence of metastases, and alkaline phosphatase (ALP) level >200 U/L. A nomogram and a 4-tier staging system predictive of OS were created using age at diagnosis, ECOG status, tumor size, presence or absence of metastasis, and ALP level. The C-statistic for this novel staging system was 0.71 compared with 0.69 for the TNM staging system (P=.08). In patients who did not undergo surgery, the C-statistics of the novel and TNM staging systems were 0.60 and 0.51, respectively (P<.001)., Conclusions: We created a novel, clinically based staging system for GBC based on nonoperative information at the time of diagnosis that was superior to the TNM staging system in predicting OS in patients who did not undergo surgery, and that performed on par with TNM staging in surgical patients.