Keeping Up With ChatGPT: Evaluating Its Recognition and Interpretation of Nuclear Medicine Images.

PURPOSE: The latest iteration of GPT4 (generative pretrained transformer) is a large multimodal model that can integrate both text and image input, but its performance with medical images has not been systematically evaluated. We studied whether ChatGPT with GPT-4V(ision) can recognize images from common nuclear medicine examinations and interpret them. PATIENTS AND METHODS: Fifteen representative images (scintigraphy, 11; PET, 4) were submitted to ChatGPT with GPT-4V(ision), both in its Default and "Advanced Data Analysis (beta)" version. ChatGPT was asked to name the type of examination and tracer, explain the findings and whether there are abnormalities. ChatGPT should also mark anatomical structures or pathological findings. The appropriateness of the responses was rated by 3 nuclear medicine physicians. RESULTS: The Default version identified the examination and the tracer correctly in the majority of the 15 cases (60% or 53%) and gave an "appropriate" description of the findings or abnormalities in 47% or 33% of cases, respectively. The Default version cannot manipulate images. "Advanced Data Analysis (beta)" failed in all tasks in >90% of cases. A "major" or "incompatible" inconsistency between 3 trials of the same prompt was observed in 73% (Default version) or 87% of cases ("Advanced Data Analysis (beta)" version). CONCLUSIONS: Although GPT-4V(ision) demonstrates preliminary capabilities in analyzing nuclear medicine images, it exhibits significant limitations, particularly in its reliability (ie, correctness, predictability, and consistency).

Prognostic Value of the De Ritis Ratio for Overall Survival in Patients with Metastatic Castration-Resistant Prostate Cancer Undergoing [177Lu]Lu-PSMA-617 Radioligand Therapy.

The De Ritis ratio (=aspartate transaminase/alanine transaminase) has shown prognostic value in different cancer types. This is the first such analysis in prostate cancer patients undergoing radioligand therapy (RLT) with [177Lu]Lu-PSMA-617. This retrospective monocentric analysis included 91 patients with a median of 3 RLT cycles (range 1-6) and median cumulative activity of 17.3 GBq. Univariable Cox regression regarding overall survival (OS) included age, different types of previous treatment, metastatic patterns and different laboratory parameters before RLT. Based on multivariable Cox regression, a prognostic score was derived. Seventy-two patients (79%) died (median follow-up in survivors: 19.8 months). A higher number of previous chemotherapy lines, the presence of liver metastases, brain metastases, a higher tumor load on PSMA-PET, a higher prostate-specific antigen (PSA) level, lower red blood cell count, lower hemoglobin, higher neutrophil-lymphocyte ratio and higher De Ritis ratio were associated with shorter OS (each p < 0.05). In multivariable Cox, a higher number of chemotherapy lines (range, 0-2; p = 0.036), brain metastases (p < 0.001), higher PSA (p = 0.004) and higher De Ritis ratio before RLT (hazard ratio, 1.27 per unit increase; p = 0.023) remained significant. This prognostic score separated five groups with a significantly different median OS ranging from 4.9 to 28.1 months (log-rank test, p < 0.001). If validated independently, the De Ritis ratio could enhance multifactorial models for OS after RLT.

Transarterial Yttrium-90 Radioembolization in Intrahepatic Cholangiocarcinoma Patients: Outcome Assessment Applying a Prognostic Score.

Radioembolization (RE) is a viable therapy option in patients with intrahepatic cholangiocarcinoma (ICC). This study delineates a prognostic score regarding overall survival (OS) after RE using routine pre-therapeutic parameters. A retrospective analysis of 39 patients (median age, 61 [range, 32−82] years; 26 females, 13 males) with ICC and 42 RE procedures was conducted. Cox regression for OS included age, ECOG, hepatic and extrahepatic tumor burden, thrombosis of the portal vein, ascites, laboratory parameters and dose reduction due to hepatopulmonary shunt. Median OS after RE was 8.0 months. Using univariable Cox, ECOG ≥ 1 (hazard ratio [HR], 3.8), AST/ALT quotient (HR, 1.86), high GGT (HR, 1.002), high CA19-9 (HR, 1.00) and dose reduction of 40% (HR, 3.8) predicted shorter OS (each p < 0.05). High albumin predicted longer OS (HR, 0.927; p = 0.045). Multivariable Cox confirmed GGT ≥ 750 [U/L] (HR, 7.84; p < 0.001), ECOG > 1 (HR, 3.76; p = 0.021), albumin ≤ 41.1 [g/L] (HR, 3.02; p = 0.006) as a three-point pre-therapeutic prognostic score. More specifically, median OS decreased from 15.3 months (0 risk factors) to 7.6 months (1 factor) or 1.8 months (≥2 factors; p < 0.001). The proposed score may aid in improved pre-therapeutic patient identification with (un-)favorable OS after RE and facilitate the balance between potential life prolongation and overaggressive patient selection.