Histologic Reaction Patterns of Lymph Node Involvement in Ovarian Serous Borderline Tumors: Their Implications for Prognosis Open Access

We identified 105 patients with SBT from the database of the Department of Pathology at Asan Medical Center, Seoul, Korea, between 2000 and 2020 whose primary tumors and LNs were surgically resected at our institution. To compare HRPs, a consecutive series of LGSCs (70 patients) were selected from patients who underwent surgical resection of the ovary and LNs during the same period at our institution. Clinical data were obtained from the electronic medical records. This study, which posed no more than minimal risk to patients, was approved by the Institutional Review Board of Asan Medical Center (approval number 2020-1697; approval date November 5, 2020), with informed consent waivers granted.

Hematoxylin and eosin–stained slides were reviewed by 2 gynecologic pathologists (J.K. and K.-R.K.), and tumor staging was determined according to International Federation of Gynecology and Obstetrics (FIGO) 2014 criteria.²³  SBT, LGSC, stromal microinvasion, micropapillary/cribriform subtype, and extraovarian implants (either invasive or noninvasive) were defined according to the 2020 World Health Organization (WHO) classification.²⁴  Cases with extraovarian “invasive implants” but with SBT in the ovarian mass were classified as LGSC instead of SBT as per the 2020 WHO classification, which has consistently classified such cases as LGSC since the 2014 edition.^24–28  LNs were examined by taking one section per node, consistent with general pathology laboratory procedures, with no additional serial sections performed based on LN size or involvement. All evaluations were conducted using hematoxylin and eosin staining, and no immunologic epithelial markers were employed.

LNI was defined as the presence of serous epithelia with low-grade cytologic atypia in forms such as single cells, cell clusters, papillary or micropapillary clusters, cribriform glands, or glands with intraglandular branching papillae.^7,10,24,25  However, the term LNI instead of LN metastasis was also used in LGSC for simplicity of comparison in the current study. Simple isolated glands lined by a single layer of tubal-type epithelium without cytologic atypia were classified as endosalpingiosis and not as LNI.^7,10 

We classified LNI into 4 categories based on tumor cell location within the afferent/efferent lymphatics, sinus, or lymphoid tissue (either cortex or medulla); the adherence of tumor cells to the surrounding lymphoid tissue; and presence or absence of a peritumoral desmoplastic reaction in the LNs. Accordingly, HRP 1 refers to tumor cells freely floating in the afferent/efferent lymphatics but not within the intranodal sinuses or LN parenchyma (Figure 1, A and E) and HRP 2 refers to tumor cells located in the subcapsular, trabecular, or medullary sinus within the LN, with clefts or empty spaces around them (Figure 1, B and F). Although free spaces were present around the tumor cells in both HRP 1 and HRP 2, we separated HRP 1 from HRP 2 because the intranodal sinus contains numerous sinus macrophages, dendritic cells, and other types of immune cells that are absent in afferent/efferent lymphatics.²⁶  HRP 3 refers to tumor cells showing firm adherence to the surrounding cortical or medullary lymphoid tissue without clefts or empty spaces around the cellular components (Figure 1, C and G), and HRP 4 refers to peritumoral desmoplastic reactions regardless of tumor cell location (Figure 1, D and H). If various types of HRPs were identified within the same or different LNs in the same patient, the highest HRP was selected. To determine the extent of LNI, the maximum linear diameter of the tumor cell clusters was recorded. If multifocal LNI was observed, the total maximum length, including all involved areas, was used to indicate the extent of involvement.

Surgery comprised either complete (bilateral salpingo-oophorectomy with or without total hysterectomy) or conservative (conserving at least a portion of one ovary) surgery. The surgical staging date was used to calculate the follow-up period. Clinical outcomes (recurrence or death) were analyzed in patients with SBT and LGSC. Recurrence of SBT was defined as histologically confirmed regrowth of the SBT or subsequent development of LGSC in the ipsilateral/contralateral ovary, LN, and/or peritoneum. Recurrence-free survival (RFS) was calculated as the duration from surgery to the first recurrence or last follow-up, excluding individuals with synchronous distant metastasis. Overall survival (OS) was defined as the duration from surgery until death from any cause or the last follow-up.

Continuous variables were compared using an independent 2-sample t test or a 1-way analysis of variance with Bonferroni adjustment. Categorical variables were compared using the χ² test or Fisher exact test, as appropriate. Kaplan-Meier survival curves were plotted and compared using the log-rank test to compare RFS and OS according to the HRP of LNI. Multivariate Cox proportional hazards regression analysis was used to identify independent prognostic variables. Statistical significance was set at P < .05, and all statistical analyses were performed using SPSS (version 21.0; IBM Corp, Armonk, New York).

The baseline clinicopathologic characteristics of patients with SBT are summarized in Table 1. Patient ages ranged from 14 to 76 years, with a mean of 41.7 ± 14.8 years. The number of dissected LNs ranged from 1 to 72 (mean, 24.8 ± 16.6), and the follow-up period ranged from 2 to 260 months (mean, 94.2 ± 65.5 months).

LNI was present in 12 of 105 (11.4%) and absent in 93 of 105 patients (88.6%). SBTs were classified as stage I, II, and III in 80 of 105 (76.2%), 6 of 105 (5.7%), and 19 of 105 cases (18.1%). None of the SBT patients presented with synchronous distant metastases. All patients with LNI were aged 50 years or younger, with a significant age difference observed between patients with and without LNI (P = .03) (Table 1). LNI was associated with bilateral ovarian involvement (P = .02), ovarian surface involvement (P = .02), and peritoneal implant (P < .001) (Table 1). A total of 9 SBT patients presented with recurrence. Four patients with FIGO stage I developed recurrence as SBT in the remnant ovary (1 patient), contralateral ovary (2 patients), and pelvis (1 patient). Two patients with FIGO stage II developed recurrence as SBT in the remnant ovary. Three patients with FIGO stage III developed recurrence as LGSC (one of whom developed recurrence 17 years after the initial surgery in the mediastinal LN).

The baseline clinicopathologic characteristics of patients with LGSC are summarized in Table 2. Patient ages ranged from 20 to 82 years, with a mean of 43.4 ± 13.4 years. The number of dissected LNs ranged from 6 to 93 (mean, 35.5 ± 18.6), and the follow-up period ranged from 3 to 242 months (mean, 67.8 ± 44.5 months). LGSC was classified as stage I, II, III, and IV in 15 of 70 (21.4%), 4 of 70 (5.7%), 45 of 70 (64.3%), and 6 of 70 cases (8.6%), and synchronous distant metastasis was present in 6 patients. LNI was associated with presence of lymphovascular invasion (P < .001) and peritoneal implant (P < .001) (Table 2).

The frequencies of LNI in patients with SBT and LGSC were 11.4% (12 of 105) and 55.7% (39 of 70), respectively, and the percentages of LNI per total dissected LNs in SBT and LGSC were 3.8% (100 of 2613) and 11.6% (289 of 2482), respectively (Table 3). This indicates that LNI was significantly less frequent in SBT than in LGSC for both the frequency of LNI (P < .001) and the percentages of LNI per total dissected LNs (P < .001; Table 3).

In SBT and LGSC, there were 100 and 289 positive LNs, respectively. Table 4 provides details regarding the number and proportion of each HRP (1–4) among the positive LNs for SBT and LGSC. LNI showing HRP 4 was significantly more frequent in patients with LGSCs (19.7%; 57 of 289) than in those with SBTs (6.0%; 6 of 100) (P = .001), whereas no significant differences in HRP 3 or higher were observed between SBTs (38.0%; 38 of 100) and LGSCs (37.7%; 109 of 289) (P = .96), thereby suggesting that peritumoral desmoplastic reactions were significantly more frequent in LGSC than in SBTs (Table 4). The extent of LNI was significantly smaller in patients with SBTs (0.7 ± 3.0 mm) than in those with LGSCs (3.4 ± 4.8 mm; P < .001; Table 3).

In SBT, HRP 3 or higher was more frequently observed in patients with peritoneal implants (P < .001) and was also more commonly associated with the micropapillary component, albeit with borderline significance (P = .06) (Table 1). HRP 4 was associated with micropapillary component (P = .02) and peritoneal implants (P = .03; Table 1). In LGSC, HRP 3 or higher was significantly associated with bilateral ovary involvement (P = .009), lymphovascular invasion (P = .008), and peritoneal implants (P < .001) (Table 2). HRP 4 also demonstrated these associations (P = .01, .01, and .007, respectively) and further correlated with age older than 50 years (P = .006) and ovary surface involvement (P = .048) (Table 2).

Overall, patients with SBT had 10- and 20-year RFS rates of 90.0 and 70.1%, respectively, and 10- and 20-year OS rates of 95.8 and 88.5%, respectively. LNI was associated with poorer RFS (P = .04) in patients with ovarian SBT, regardless of the presence or absence of extraovarian implants (Figure 2, A). In SBT, HRP 3 or higher was associated with poorer RFS (P = .02; Figure 2, B). Moreover, HRP 4 was significantly associated with decreased RFS in patients with SBT (P = .04; Figure 2, C). An LNI extent of 1 mm or more was associated with a poorer RFS (P = .04; Figure 2, D). However, OS was not affected by the presence or absence of LNI, HRP 3 or higher, HRP 4, or LNI extent of 1 mm or more (P = .34, .32, .21, and .34; Supplemental Figure 1, A through D; see supplemental digital content file at https://meridian.allenpress.com/aplm in the July 2025 table of contents containing 2 figures). To determine the independent impact of HRPs on RFS, we adjusted for potential confounding factors related to the primary tumor, including the micropapillary component, and nodal tumor characteristics, including the maximal extent of LNI (mm), in a multivariate Cox regression analysis. The presence of HRP 3 or higher significantly influences RFS, with a hazard ratio of 6.93 (95% CI, 1.15–41.88; P = .03), independent of micropapillary component and maximal extent of LNI (Table 5). However, HRP 4, with a hazard ratio of 5.58 (95% CI, 0.90–34.53; P = .06), was not independently significant for RFS (Table 5). Complementary findings in Supplemental Figure 2 further illustrate that LGSC patients with LNI exhibited worse RFS and OS than those without LNI, with HRP 3 or higher and HRP 4 specifically associated with more severe declines in RFS, indicating a robust correlation between HRPs and survival outcomes.

In SBT, no patients with HRP 1 or 2 experienced tumor recurrence, whereas those with HRP 3 or higher demonstrated reduced RFS (Figure 3, A). This contrasts sharply with the outcomes observed in LGSC, where patients with HRP 1 or 2 also experienced tumor recurrence, occurring at a rate similar to those observed in patients with HRP 3 or higher (Figure 3, B). This finding underscores the indolent nature of HRP 1 and HRP 2 especially in SBT, in contrast to HRP 3 or higher. Given the sharp contrast between HRP 1 or 2 and HRP 3 or higher in SBT, unlike in LGSC, these results highlight the distinct tumor biology inherent to SBT.

In the current study, we assessed the prognostic significance of HRP in 105 patients with SBT, making it one of the most extensive studies on the relevance of LNI in patients with SBT. Our findings suggest that peritumoral HRP in LNs, which indicates firm adhesion of tumor cells to the surrounding lymphoid tissue with or without a desmoplastic reaction (HRP ≥3), is associated with poorer RFS in patients with SBT, as confirmed by our multivariate analysis.

Based on current literature, it is suggested that adverse prognostic outcomes in patients with HPR 3 or higher can be attributed to several factors. First, CAFs facilitate desmoplastic reactions around tumor cells, promoting tumor growth, migration, and invasion by altering the tumor microenvironment, leading to tumor progression, metastasis, and treatment resistance.^18–21  Nodal desmoplasia may foster a tumor-permissive immunologic microenvironment by activating CAFs and promoting T-helper 2 cell infiltration while also acting as a barrier to immune cell migration toward tumor cells.^27,28  Furthermore, the firm adherence of tumor cells to LN parenchyma without clefts or empty spaces in HRP 3 or higher could restrict the access of nodal immune cells, impeding antitumor immunity. Sinus macrophages directly clasp tumor cells entering from the afferent lymphatic vessels, preventing their invasion into the nodal parenchyma or systemic spread.²⁶  Thus, LNI with HRP 3 or higher may limit the access of sinus macrophages or dendritic cells to tumor cells and disrupt frontline immune defense.²⁶  Moreover, the higher frequency of HRP 3 or higher in LGSC-positive LNs than that in SBT and LGSC implies that primary tumors with HRP 3 or higher tend to be more aggressive or prone to invasiveness. This association suggests that firm adhesion or intranodal desmoplasia, as indicated by HRP 3 or higher, may reflect a more aggressive nature in primary tumors. Notably, within the SBT cohort, cases with HRP 3 or higher were associated with more aggressive features, including the presence of peritoneal implants and a borderline association with the micropapillary component.

Although one study has linked nodular aggregates of tumor cells larger than 1 mm within LNs to reduced disease-free survival,⁷  most studies found that LNI could not be deemed an independent prognostic factor for SBT recurrence^{1,3,6,8,10–14}  or mortality.^{1,3,5,7,9,10,12–14}  Although HRP 3 or higher in our study and “nodular aggregates larger than 1 mm” identified by McKenney et al⁷  are both linked to increased recurrence, our approach includes a broader range of tumor growth patterns, including nodular aggregates, single cells, cell clusters, papillae, and glandular/intraglandular proliferation, within HRP 3 or higher. Furthermore, HRP 3 or higher emphasizes the firm adhesion of tumor cells to surrounding lymphoid tissue, highlighting the interplay between tumor cells and the surrounding normal lymphoid tissue. In SBT, patients with HRP 1 or 2 did not experience any tumor recurrence, underscoring the nonrecurrent nature of these HRPs. On the other hand, those with HRP 3 or higher demonstrated a significant reduction in RFS, highlighting potential aggressiveness of these patterns. This pattern suggests that the worse prognosis in patients with LNI-positive SBT could be primarily associated with the firm adhesion of tumor cells to surrounding tissue (HRP ≥3) or peritumoral desmoplastic reaction (HRP 4).

Although LGSC cases showed recurrence across all HRPs, including HRP 1 or 2, this contrast emphasizes the unique biological behavior of SBT. We propose that the presence of HRP 1 or 2 in SBT does not represent true metastatic involvement, in contrast to LGSC, and therefore these patterns should not be included in the total count of positive LNI in SBT. Floating tumor cells, predominantly present in the afferent/efferent lymphatic and/or subcapsular sinus (HRP 1–2), appeared in approximately two-thirds (62 of 100; 62.0%) of the involved LNs of SBT. The buttonlike structure of the lymphatic endothelial junction, which consists of overlapping endothelial cells that create a unique seal, facilitates the selective transport of cells and fluids into the lymphatics and is crucial for clarifying how tumor cells are transported from the primary ovarian mass to the afferent lymphatics and the subcapsular sinus.^2,29  Conversely, HRP 3 or higher may denote an impending invasion or early invasive stage of tumor cells within LNs, akin to invasive implants.

Ovarian SBTs with invasive implants, currently classified as LGSC owing to their invasive growth and desmoplastic histologic reaction, were diagnosed as SBT in earlier decades. Early LNI prognostic studies conducted pre–2014 WHO classification^* may have included such cases, leading to variable results. Our current WHO-compliant study²⁴  indicates that LNI-positive SBT correlates with reduced RFS.

There are several limitations to our study that should be considered. First, the sample size of SBT cases with LNI was relatively small (12 cases), which may limit the generalizability of our findings. Additionally, the retrospective nature of the study and the reliance on existing pathologic data may have introduced selection bias. Further prospective studies with larger cohorts are needed to validate our findings and provide more definitive conclusions regarding the prognostic significance of HRP in SBT.

In conclusion, our study revealed that LNI in SBT is associated with decreased RFS, particularly when linked to HRPs characterized by firm adhesion of tumor cells to surrounding tissue, with or without a desmoplastic reaction (HRP ≥3). This pattern was found to independently impact RFS negatively. Conversely, LNI associated with freely floating tumor cells in afferent/efferent lymphatics or intranodal sinuses, which feature surrounding free spaces (HRP 1–2), did not exhibit any recurrence. These findings underscore the importance of differentiating between these HRPs in assessing the prognosis and managing patients with SBT.