Cervical cancer incidence worldwide is expected to decline with the widespread adoption of human papillomavirus (HPV) vaccination; however, the trend is not uniform across all regions. In Japan, the incidence of cervical cancer among women under 50s is on the rise, and given the historically low HPV vaccination rate, this increase is projected to continue [[1], [2], [3]].

Cervical adenocarcinoma accounts for a significant minority of cervical cancers, and its incidence appears to be increasing, both in proportion to cervical cancer overall and relative to the general population [4]. Compared with squamous cell carcinoma, adenocarcinoma is considered less radiosensitive, prompting some gynecologists to favor surgical treatment over radiotherapy for early-stage disease [5]. Surgical approaches vary widely, from conization to radical hysterectomy with lymphadenectomy, each carrying different degrees of morbidity. Accurate patient stratification is therefore critical to minimize overtreatment and reduce treatment-related harm [6].

Current treatment recommendations for cervical adenocarcinoma (and cervical cancer more broadly) are primarily based on tumor stage, [5,7,8] which for early-stage disease is dictated by tumor size and depth of invasion. Histological tumor characteristics play a limited role in guiding treatment decisions, with the notable exception of lymphovascular space invasion (LVSI) as included in the Sedlis criteria [9]. However, this stage-based framework may not adequately capture the biological heterogeneity of cervical adenocarcinomas.

The latest WHO classification of female genital tract tumors recognizes HPV-associated (HPVA) and HPV-independent (HPVI) types of cervical adenocarcinoma [[10], [11], [12]]. This classification reflects growing recognition of HPVI subtypes in recent decades, particularly gastric-type adenocarcinoma, [13] which is associated with worse clinical outcomes than HPVA [[14], [15], [16], [17], [18]]. Some guidelines exclude them from candidates for fertility-sparing treatment [5,8]. While the prevalence of HPVI varies by region, its relative importance is expected to increase in countries with widespread HPV vaccination.

Efforts to further classify HPVA, which remains the majority subtype, are also underway. The Silva classification was developed primarily to identify patients with a negligible risk of metastasis who may be eligible for conservative management [19]. This system categorizes HPVA into three stromal invasion patterns (A, B, and C). Pattern A tumors exhibit extremely low malignant potential and are rarely associated with lymph node metastasis, suggesting that strict distinction from adenocarcinoma in situ (AIS) may be unnecessary in clinical practice. The system also simplifies diagnostic interpretation for pathologists, who have historically struggled to define stromal invasion in glandular lesions. A more recent development is the binary Silva classification, which combines stromal invasion patterns with the presence or absence of LVSI. This system groups pattern A and pattern B without LVSI into a low-risk category, and pattern B with LVSI and pattern C into a high-risk category. It aims to improve diagnostic reproducibility [[20], [21], [22]] while maintaining clinical relevance, [23] and emerging molecular data suggest that these two groups may be biologically distinct [24].

Despite these developments, the clinical use of Silva classifications remains limited, and their impact on treatment decisions has yet to be realized. This situation is likely due to the lack of sufficient high-quality studies to firmly establish the clinical utility of Silva classifications. The key studies on the Silva classification [19,[25], [26], [27]] have been international efforts that compiled cases from around the world. While such studies may enhance the generalizability of pathological diagnosis, they likely suffer from heterogeneity in treatment protocols. Moreover, some prior studies did not distinguish between HPVA and HPVI [23,28,29]. Limited evidence suggests that HPVI may be even more aggressive than HPVA with invasion pattern C, [29,30] but the question of how HPVI should be positioned in relation to HPVA classified by the Silva systems remains largely unexplored. The validation of the binary Silva classification is still in its early stage, with only two studies to date having analyzed cohorts comprising several hundred cases [23,27]. More evidence from cohorts managed under consistent clinical protocols is needed to establish its prognostic value.

This multicenter retrospective study was conducted to validate the risk stratification of HPVA based on the Silva classifications and to clarify how HPVI should be positioned in relation to HPVA subgroups defined by this system. Notable features of this study include its restriction to surgically treated cases, participation limited to Japanese institutions, and centralized review of MRI images and histological slides. Our findings contribute to the growing body of evidence supporting the clinical utility of histomorphological classification in cervical adenocarcinoma and provide further justification for its implementation in clinical practice.