Patients with pT1 high-grade (HG) urothelial carcinoma (UC) and a very high risk of progression might benefit from immediate radical cystectomy (RC), but this option remains controversial. Validation of a standardized method to evaluate the extent of lamina propria (LP) invasion (with recognized prognostic value) in transurethral resection (TURBT) specimens is still needed. The Rete Oncologica Lombarda (ROL) system showed a high predictive value for progression after TURBT in recent retrospective studies. The ROL system was supposed to be validated on a large prospective series of primary urothelial carcinomas from a single institution. From 2016 to 2020, we adopted ROL for all patients with pT1 HG UC on TURBT. We employed a 1.0-mm threshold to stratify tumors in ROL1 and ROL2. A total of 222 pT1 HG UC were analyzed. The median age was 74 years, with a predominance of men (73.8%). ROL was feasible in all cases: 91 cases were ROL1 (41%), and 131 were ROL2 (59%). At a median follow-up of 26.9 months (IQR 13.8–40.6), we registered 81 recurrences and 40 progressions. ROL was a significant predictor of tumor progression in both univariable (HR 3.53; CI 95% 1.56–7.99; p < 0.01) and multivariable (HR 2.88; CI 95% 1.24–6.66; p = 0.01) Cox regression analyses. At Kaplan-Meier estimates, ROL showed a correlation with both PFS (p = 0.0012) and RFS (p = 0.0167). Our results confirmed the strong predictive value of ROL for progression in a large prospective series. We encourage the application of ROL for reporting the extent of LP invasion, substaging T1 HG UC, and improving risk tables for urological decision-making.

In recent years, immunohistochemical protein expression was studied as a surrogate to the molecular classification of bladder cancer, although no tissue biomarkers are available for clinical use to predict survival or the response to neoadjuvant chemotherapy (CT) in UC, as the literature produced conflicting results. This retrospective study included TURB specimens harboring foci of HG pT2 muscle-invasive bladder carcinoma (MIBC) from 251 patients who subsequently underwent radical cystectomy. We performed immunohistochemical analysis on tumor samples, for relevant gene-expression-based markers for basal type (CD44, CK5/6) and luminal type (CK20 and pPARγ). Piescore, investigated in both non-muscle-invasive (NMI) and muscle-invasive (MI) components of the tumor, divided basal and luminal UC-types when at least three of the four markers were consistent with a specific phenotype, mixed types if one/two luminal and basal markers were present simultaneously, and neu-like types when all four markers investigated were negative. Eighteen selected cases were also investigated with RT-PCR to validate, and to increase the specificity of, the immunohistochemical results. We observe an immunophenotypical difference in the NMI and MI components in 96/251 UC patients (38.25%): half of tumors (44/96 cases) have a transition to basal, 36.46% (35/96 cases) to neu-like, 12.5% (12/96 cases) to mixed, and 5.2% (5/96 cases) to luminal phenotypes. Mixed tumors in the NMI component are more likely to change phenotype than other groups, particularly compared with basal tumors, which demonstrate greater stability (only 8/96 cases, p < 0.00001). The transition of luminal tumors to basal display a better OS compared with the transition toward neu-like tumors (p = 0.027). Overall, the phenotypical switch does not affect lymphovascular invasion, pT, DFS, or OS compared with non-switched cases. In the MI component, the presence of CD44 expression, irrespective of score-related phenotype, shows a protective effect in papillary-type UC (OS p = 0.008, HR 0.453, PFS p = 0.07, HR 0.599), and in UC naïve for CT (p = 0.0479). Piescore immunophenotyping reveals an intratumoral phenotypical transition between the NMI and MI components of the same tumor. The molecular change is a common event in the mixed and luminal categories, but not in basal tumors, which show better phenotypical stability. This phenomenon could partially explain the sensitivity of a subset of luminal UC to chemotherapy: good responders could be “non-real” luminal UC, which acquire nasal markers, such as CD44.

ChatGPT is an artificial intelligence capable of processing and generating human-like language. ChatGPT’s role within clinical patient care and medical education has been explored; however, assessment of its potential in supporting histopathological diagnosis is lacking. In this study, we assessed ChatGPT’s reliability in addressing pathology-related diagnostic questions across 10 subspecialties, as well as its ability to provide scientific references. We created five clinico-pathological scenarios for each subspecialty, posed to ChatGPT as open-ended or multiple-choice questions. Each question either asked for scientific references or not. Outputs were assessed by six pathologists according to: 1) usefulness in supporting the diagnosis and 2) absolute number of errors. All references were manually verified. We used directed acyclic graphs and structural causal models to determine the effect of each scenario type, field, question modality and pathologist evaluation. Overall, we yielded 894 evaluations. ChatGPT provided useful answers in 62.2% of cases. 32.1% of outputs contained no errors, while the remaining contained at least one error (maximum 18). ChatGPT provided 214 bibliographic references: 70.1% were correct, 12.1% were inaccurate and 17.8% did not correspond to a publication. Scenario variability had the greatest impact on ratings, followed by prompting strategy. Finally, latent knowledge across the fields showed minimal variation. In conclusion, ChatGPT provided useful responses in one-third of cases, but the number of errors and variability highlight that it is not yet adequate for everyday diagnostic practice and should be used with discretion as a support tool. The lack of thoroughness in providing references also suggests caution should be employed even when used as a self-learning tool. It is essential to recognize the irreplaceable role of human experts in synthesizing images, clinical data and experience for the intricate task of histopathological diagnosis.