Identification of low-risk cases of invasive endocervical adenocarcinoma with Silva pattern-based classification: a systematic review and meta-analysis

Abstract

Objective

We aimed to describe the discriminatory ability of the Silva pattern-based classification system for invasive endocervical adenocarcinoma (EAC) in predicting risks of lymph node (LN) metastasis, recurrence, and death.

Method

We systematically searched PubMed, Scopus, and Embase through 2024 for manuscripts describing patients with EAC by Silva pattern-based classification. We included studies reporting outcomes of LN metastasis, recurrence, or death by Silva pattern. Random-effects meta-analysis was used to summarize binomial proportions and compare outcomes between groups.

Results

We identified 19 studies including 2998 patients (20.8 % pattern A, 23.1 % pattern B, 54.1 % pattern C). Silva pattern A cases showed a significantly lower risk of LN metastasis (3/509; 0.6 %) than pattern B (6.1 %; OR = 0.33, 95 %CI 0.16–0.67) or pattern C (22.0 %; OR = 0.09; 95 %CI 0.05–0.16). Across 5 studies limited to stage I disease, there were no Silva pattern A cases with LN metastasis, recurrence, or death. Among 11 studies reporting on recurrence and death, Silva pattern A cases had distinctly low risk of 0.3 % for both outcomes, including a significantly lower risk of recurrence (OR = 0.15, 95 %CI 0.06–0.34) and death (OR = 0.21, 95 %CI 0.09–0.50) versus patterns B/C. Silva pattern B cases showed significantly lower risk of all outcomes compared with pattern C.

Conclusions

We demonstrate that Silva classification is highly predictive of oncologic outcomes for EAC, with patterns A, B, and C showing distinct low, intermediate, and high risk profiles. The low risk profile of pattern A may justify future prospective study of surgical and adjuvant treatment de-escalation from the current standard of care, while pattern C cases may warrant consideration of adjuvant treatment escalation.