[Utility of digital pathology image analysis in micropapillary serous borderline ovarian tumor]

Objective: To investigate the potential value of quantitative digital pathology analysis in predicting prognosis in micropapillary serous borderline ovarian tumor (MP-SBOT). Methods: Clinical and pathological data from 366 serous borderline ovarian tumor (BOT) patients who attended the First Affiliated Hospital of Nanjing Medical University between October 2012 and November 2023 were retrospectively reviewed. Patients were classified into MP-SBOT group and non-MP-SBOT group according to their pathological subtype. Within the MP-SBOT group, patients were further classified into recurrence and non-recurrence subgroups. Digital image analysis using QuPath was performed on 83 ovarian serous tumors with adequate tissue quality, including MP-SBOT (n=48), SBOT in the non-MP-SBOT group (n=9), serous cystadenoma (n=11), low-grade serous carcinoma (n=7) and high-grade serous carcinoma (n=8). Quantitative pathological features including nuclear area, cellular area, and nuclear-to-cytoplasmic ratio (N/C ratio) of tumor cells were extracted. Univariate and multivariate Cox regression analyses were used to identify factors associated with MP-SBOT recurrence. Kaplan-Meier method was applied to estimate progression-free survival (PFS) time. Results: (1) Compared with non-MP-SBOT group, MP-SBOT group patients exhibited higher median serum carbohydrate antigen 125 level (127.9 vs 36.5 kU/L), higher D-Dimer level (0.7 vs 0.3mg/L), higher proportion of bilateral ovarian tumors (60.4% vs 15.4%), higher proportion of invasive peritoneal implants (25.0% vs 3.1%), higher proportion of International Federation of Gynecology and Obstetrics (FIGO) stage Ⅱ-Ⅲ (54.1% vs 10.0%), higher proportion of mixed cystic-solid mass detected by imaging examination (88.1% vs 70.1%) and higher proportion of recurrence (33.3% vs 13.5%) with statistically significant differences (all P<0.001). (2) Quantitative digital pathology revealed the distinct cellular morphologic characteristics in MP-SBOT. The tumor cell N/C ratio was significantly higher in MP-SBOT group than non-MP-SBOT group (0.48 vs 0.41; P<0.05). The tumor cell N/C ratio increased progressively with tumor malignancy. (3) Univariate Cox analysis identified tumor cell N/C ratio>0.50 and conservative surgery as factors associated with recurrence in MP-SBOT group. Multivariate analysis confirmed that an elevated tumor cell N/C ratio was an independent risk factor for recurrence in MP-SBOT group (HR=7.278, 95%CI:1.940-27.307, P=0.003). (4) Kaplan-Meier survival analysis showed that patients with tumor cell N/C ratio>0.50 had a significantly shorter median PFS than those with the tumor cell N/C ratio≤0.50 (164.0 vs 35.0 months, P<0.001). Conclusions: MP-SBOT display more aggressive clinical features. The elevated tumor cell N/C ratio shows potential clinical value in predicting recurrence in MP-SBOT patients.