Trevor Starnes, M.D., M.P.H., Surgical Pathology Fellow
John Bishop, M.D., Director of Anatomic Pathology
Anthony Karnezis, M.D., Ph.D., Associate Professor, Gynecologic Surgical Pathology
In economically developed countries like the United States, the decline of endocervical squamous cell carcinoma associated with the utilization of the Pap screening test is one of the great success stories in medicine. However, at the same time, there has been a relative and absolute increase in endocervical adenocarcinomas so that they comprise 25% of all cervical cancers (Adegoke, 2012). Though not as reliably associated with human papillomavirus (HPV) infection as squamous cell carcinoma, about 85% are HPV-associated. HPV strain 18 accounts for half of these.
As with most cancers, evidence-based staging is critical to balance the risks and benefits of treatments for these cases. The current 2018 FIGO staging of cervical carcinomas relies upon depth of stromal invasion for early-stage cancers (IA1-IB1), then macroscopic tumor size (IB2-IB3), locally advanced spread (IIA1-IIIB), nodal metastasis (IIIC1-IIIC2), and finally extension beyond the true pelvis or to the mucosal surfaces of bladder or rectum (IVA-IVB) (Bhatla et al, 2018). The pathologist’s contribution to early staging relies upon measuring stromal invasion.
Overall, this system makes sense and correlates with patient outcomes. One difficulty with using stromal invasion for endocervical adenocarcinoma is that the point where invasion begins is not always clearly defined. Normal endocervical glands can dive deep into stroma, and when present are used to determine where invasion begins. A <1 mm difference in invasion can change the stage and thus how aggressive treatment will be. Fortunately, supplemental information can be helpful in treating each patient’s case individually. An already-appreciated example is the presence of lymphovascular invasion. A less well-known factor is pattern of invasion.
The Silva system
A 2013 article proposed a way to risk-stratify endocervical adenocarcinomas based upon the pattern of invasion (De Vivar et al., 2013). It is known as the Silva system. They used three tiers from indolent to aggressive as Pattern A, Pattern B, and Pattern C. Pattern A is present when all glands lack destructive stromal invasion, do not exhibit solid growth, and have no lymphovascular invasion. Pattern C is diffuse destructive stromal invasion with a desmoplastic response, or at least 5mm2 of confluent growth. Pattern B is in between, but also without solid growth. In their cases, approximately 21% are Pattern A.
Figure 1: Left: Pattern A with a pushing border but no destructive invasion.
Pattern B with foci of destructive invasion with desmoplastic reaction
Figure 2: Pattern C with diffuse destructive stromal invasion
The Silva system seems intuitive, and results were compelling when this system was applied to their study population. They used a retrospective analysis of 352 women who had endocervical adenocarcinoma of the usual type, a resection with negative margins and at least two lymph nodes. Of these, 73 women (20.7%) had Pattern A. Of patients with Pattern A, none had positive lymph nodes or recurrences in 1,333 lymph nodes sampled. In contrast, Pattern C patients had positive lymph nodes 24% of the time, a 22% recurrence rate, and a 9% disease-related mortality rate.
The average depth of cervical stromal invasion was lower with Pattern A, but there were exceptions. About 27% of Pattern A cases had invasion greater than 5 mm, while 26% of Pattern C cases had invasion of less than 5 mm (De Vivar et al, 2013). A follow-up article noted that one Stage IA1 but Pattern C case had a lymph node metastases on resection (7% of Stage IA, Pattern C cases). Though only one case, it does suggest that Pattern C can be dangerous even with shallow invasion (Roma et al, 2016), while the same was not demonstrated for Pattern A even with deeper invasion.
Table 1: Outcome table from the paper by De Vivar et al, 2013
More recently, other research groups have corroborated the Silva system. For example, a group from China examined their endocervical adenocarcinomas retrospectively and found 213 cases. While a smaller proportion of them were Pattern A (13%), they had no lymph node metastases in Pattern A or B cases with 772 lymph nodes examined, in contrast to 26% of Pattern C cases with lymph node metastases (Xu et al, 2019).
A group in the Netherlands applied the system to their 82 cases (Spaans et al 2018). They also noted no recurrences or lymph node metastases in Pattern A patients, and a 35% recurrence rate for Pattern C. They observed a lower mutation burden for Pattern A cases, with a minority positive for KRAS mutations but consistently lacking mutations in PIK3CA, PTEN, and others found in Patterns B and C.
A multinational research group known for their work with non-HPV related endocervical adenocarcinomas applied the system to their cohort which included 292 HPV-associated endocervical adenocarcinomas (Stolnicu et al, 2018). They noted no lymph node metastases in both Pattern A and Pattern B cases, but that the system could not be applied to the approximately 15% of endocervical adenocarcinomas which are not HPV-related, such as the gastric, clear cell, and endometrioid subtypes.
The evidence is promising that the Silva system is helpful in identifying cases that may warrant a different treatment approach than depth of invasion alone suggests. The caveat is that the studies are all retrospective, and no clinical trial has yet been completed. Yet with the consistent result of no recurrences or lymph node metastases when a pure Pattern A is observed, this system could be helpful in this situation today. It’s important to note that assigning Pattern A still requires a full excision of the tumor. So, while a small biopsy may show no evidence of Pattern C – designating the case as a Pattern A would require a cone biopsy with negative margins at a minimum as focal destructive stromal invasion, or lymphovascular invasion, would render it Pattern B.
While potentially lifesaving in aggressive cancers, radical hysterectomies, pelvic lymph node dissections, and pelvic radiotherapy each have risks. A patient with a fully resected, Pattern A endocervical adenocarcinoma could avoid the potential complications of further treatments with a very low risk of recurrence. Such a recurrence has not yet been documented in multiple studies.
After analyzing the literature mentioned above, we are convinced that pathologists ought to start reporting whether destructive stromal invasion is present in endocervical adenocarcinoma cases to give clinicians a valuable piece of supplemental information for treatment decisions. Perhaps one day clinical trials will also support an improved staging system incorporating pattern of invasion.
Adegoke et al. “Cervical cancer trends in the United States: a 35-year population-based analysis.” Journal of Womens Health. 2012 Oct; 21(10): 1031–1037.
Bhatla N et al. “Revised FIGO staging for carcinoma of the cervix uteri” Int J Gynecol Obstet. 2019; 145: 129-135.
De Vivar A et al. “Invasive Endocervical Adenocarcinoma: Proposal for a New Pattern-based Classification System With Significant Clinical Implications: A Multi-institutional Study” Int J Gynecol Pathology. 2013; 32(6): 592-601.
Park K and Roma A. “Pattern based classification of endocervical adenocarcinoma: a review” Pathology. 2018; 50(2): 134-140.
Roma et al. “New pattern-based personalized risk stratification system for endocervical adenocarcinoma with important clinical implications and surgical outcome” Gynecologic Oncology. 2016; 141: 36-42.
Spaans V et al. “Independent validation of the prognostic significance of invasion patterns in endocervical adenocarcinoma: Pattern A predicts excellent survival” Gynecologic Oncology. 2018; 151: 196-201.
Stolnicu S et al. “Stromal invasion pattern identifies patients at lowest risk of lymph node metastasis in HPV-associated endocervical adenocarcinomas, but is irrelevant in adenocarcinomas unassociated with HPV” Gynecologic Oncology. 2018; 150: 56-60.
Xu H et al. “Expanded study on the risk of lymphovascular space invasion and lymph node metastasis of endocervical adenocarcinoma using Pattern Classification: a single-centre analysis of 213 cases” Pathology. 2019 Oct;51(6): 570-578.