Objective: To directly compare the diagnostic accuracy of deep learning models with human experts and other diagnostic methods used for the clinical detection of oral cancer.Methods: Comparative diagnostic studies involving patients with photographic images of oral mucosal lesions (cancer or non‐cancer) were included. Only studies using deep learning methods were eligible. Medline, EMBASE, Scopus, Google Scholar, and ClinicalTrials.gov were searched until September 2024. QUADAS‐C assessed the risk of bias. A Bayesian meta‐analysis compared diagnostic test accuracy.Results: Eight studies were included, none of which had a low risk of bias. Three studies compared deep learning versus human experts. The difference in sensitivity favored deep learning by 0.024 (95% CI: −0.093, 0.206), while the difference in specificity favored human experts by −0.041 (95% CI: −0.218, 0.038). Two studies compared deep learning versus postgraduate medical students. The differences in sensitivity and specificity favored deep learning by 0.108 (95% CI: −0.038, 0.324) and by 0.010 (95% CI: −0.119, 0.111), respectively. Both comparisons provided low‐level evidence.Conclusions: Deep learning models showed comparable sensitivity and specificity to human experts. These models outperformed postgraduate medical students in terms of sensitivity. Prospective clinical trials are needed to evaluate the real‐world performance of deep learning models.

Background: Recurrent oral squamous cell carcinoma (rOSCC) is associated with a poor prognosis, and evidence supporting risk stratification in the salvage setting remains limited. The prognostic value of conventional staging systems and tumor-related characteristics after recurrence has not yet been fully established. Methods: We conducted a single-center observational study including patients with rOSCC who had previously undergone treatment for pT1–2 N0M0 OSCC, all subsequently surgically salvaged. Clinical, histopathological, and recurrence-related variables were collected. Recurrent tumors were classified according to the AJCC TNM staging system, the Composite Oral Cavity and Oropharynx Recurrence Staging System (CORSS), and the Same–Up–Down staging system (SUDss), a novel staging system developed by our group. Results: A total of 49 patients were included. Several pathological characteristics, including the rT3–4 category (p = 0.025), perineural invasion (p < 0.001), lymphovascular invasion (p = 0.003), greater depth of invasion (p = 0.011), nodal involvement (p = 0.002), and higher rTNM (p = 0.004) and CORSS stage (stage II p = 0.012; stage III p = 0.038) were associated with poorer OS in univariable analyses. In multivariable analysis, CORSS stage II (HR 2.89, 95% CI 1.04–8.04; p = 0.041) and stage III (HR 12.73, 95% CI 2.11–76.68; p = 0.005), as well as close/positive surgical margins (HR 2.57, 95% CI 0.99–6.71; p = 0.045) and pathological nodal involvement (HR 4.82, 95% CI 1.77–13.14; p = 0.002), were independently associated with poorer OS. Conclusions: CORSS appears to offer complementary prognostic value compared with conventional rTNM in patients with surgically treated rOSCC, although formal comparative validation of discrimination and calibration is warranted. Pathological lymph node involvement and close or positive surgical margins were the most relevant independent adverse prognostic factors. These findings identify higher-risk patients but do not, by themselves, establish a benefit from treatment intensification.