The risk of lymph node metastasis in the new FIGO 2018 stage IA cervical cancer with >7 mm diameter

1. Design and setting

A retrospective cohort study was performed by analyzing data from the electronic patient records of the Amsterdam University Medical Center (UMC) and the University Medical Center Groningen (UMCG). All consecutive patients, newly diagnosed with FIGO (2009) IB cervical cancer between February 1985 and August 2020 and treated by (open) radical hysterectomy, radical trachelectomy, simple hysterectomy, cervical conization or large loop excision of the transformation zone (LLETZ), combined with a complete pelvic lymphadenectomy, were identified. In case of LNM, parametrial involvement or tumor in the resection margin, adjuvant (chemo)radiotherapy was recommended. At the UMCG the Sedlis criteria were also used to recommend adjuvant radiotherapy [8]. All patients with an adenocarcinoma, squamous cell carcinoma, or adenosquamous cell carcinoma of the cervix, with ≤5 mm DOI and >7 mm diameter were included in the study. If information regarding clinical visibility of the lesion was available, patients were further subdivided into a group in which the diagnosis was only made by microscopy, and another group in which the tumor was clinically visible. Patients were excluded if no hysterectomy was performed, no lymph nodes were examined, the DOI was unknown, or if neoadjuvant chemotherapy was administered.

2. Ethical approval

Ethical approval for the study was obtained from the Institutional Review Board (IRB) of the Amsterdam UMC (reference number W20_448 # 20.495). Written informed consent was waived by the IRB, because, according to Dutch law, this is not obligated in case anonymized patient data are used, keeping in mind the rules of good clinical practice.

3 Data collection

The medical records of the included patients were reviewed. The following data were retrieved from these records: age at diagnosis, clinical visibility of the lesion, primary treatment, date of surgery, adjuvant therapy, last follow-up date, date of recurrence, and date of death. A tumor was considered clinically visible when a suspect lesion was seen at routine gynecological examination.

Pathology reports were evaluated, and the following data were retrieved: histological subtype, diameter (in millimeters), DOI (in millimeters), presence of positive lymph nodes, and presence of LVSI. The Amsterdam UMC pathologic specimens from 1985 till 1992 were previously reviewed by an expert gynecologic pathologist [9], whereas pathologic specimens from 1992 till 2020 were analyzed prospectively by expert gynecologic pathologists and discussed in a multidisciplinary board. All UMCG pathologic specimens were analyzed by an expert gynecologic pathologist. For this study, all specimens of patients with positive lymph nodes were reviewed again by an expert pathologist (J.B; M.C.G.B.). If any of these cases turned out to have a DOI of >5 mm, they were excluded from the study. Four cases with LNM were not revised since there were no specimens available. In this case, the DOI, diameter and LVSI described in the original report was used.

4. Statistical analysis

Descriptive statistics were used to summarize the basic demographic details. The incidence of LNM was calculated with 95% confidence interval (CI) for the whole population. The associations of DOI, diameter, histological type, clinical visibility of the lesion and LVSI with LNM were evaluated by calculating the odds ratios (OR) with 95% CI using logistic regression. P-values <0.05 were considered statistically significant. IBM SPSS Statistics 26 (IBM Corp., Armonk, NY, USA) was used for this study. In accordance with the journal’s guidelines, we will provide our data for the reproducibility of this study in other centers if such is requested.

1. Patient selection

A total of 1,291 patients with 2009 FIGO stage IB were identified, of which 389 (30.1%) patients had a DOI ≤5 mm and diameter >7 mm (Fig. 1). In 265 of 389 patients a description of whether the tumor was seen by microscopy only or was clinically visible was found in the medical records.

Fig. 1
Patient selection.
FIGO, International Federation of Gynecology and Obstetrics; LNM, lymph node metastasis.

• Click for larger image
• Download as PowerPoint slide

Pathology of tumor specimens of 19 patients with positive lymph nodes were reviewed. Nine of nineteen reviewed cases turned out to have a DOI of >5 mm, these cases were excluded. Although described as LVSI negative in the original report, three reviewed cases turned out to have LVSI.

2. Patient and disease-related characteristics

The median age at diagnosis was 41 years (range 23–86 years). The median follow-up duration was 62 months. Patient characteristics are shown in Table 1. The most common primary treatment was a radical hysterectomy with pelvic lymphadenectomy (93.8%). Some patients received a simple trachelectomy (3.3%), simple hysterectomy (0.8%) or conization or re-LLETZ (2.1%) with pelvic lymphadenectomy. Of all patients, 3.6% received adjuvant radiotherapy and 0.5% patients received adjuvant chemoradiation. The five-year recurrence free survival was 97.7% (95% CI=95.7%–98.8%). The five-year survival rate was 97.7% (95% CI=95.7%–98.8%).

Table 1
Patient characteristics

• Click for larger image
• Click for full table
• Download as Excel file

Table 2 shows the disease-related characteristics. The most prevalent histological subtype was squamous cell carcinoma (68.6%). All patients had a DOI of ≤5 mm, but in 18 cases the exact DOI (≤3 mm or >3 mm) was unknown. LVSI was found in 23.9% of the patients. In 38.9% cases the lesion was seen by microscopy only.

Table 2
Disease-related characteristics

• Click for larger image
• Click for full table
• Download as Excel file

3. Lymph node metastasis

The incidence of pathologically confirmed LNM was 2.6% (95% CI=1.3%–4.5%). Figs. S1 and S2 show how the cases with LNM are distributed among different tumor characteristics groups. No difference in risk of =LNM was found between tumors with a DOI of ≤3 mm (0.6%) and tumors with a DOI of >3 and ≤5 mm (3.9%, p=0.074). The incidence of LNM was similar in the different histological subtypes. No cases with LNM were found in the tumors with a DOI of ≤3 mm without LVSI.

In the univariate analysis, the presence of LVSI increased the risk of LNM (OR=7.95; 95% CI=2.01–31.40). Other characteristics were not associated with LNM (Table 3).

Table 3
Univariate analysis lymph node metastases

• Click for larger image
• Click for full table
• Download as Excel file

Of the 34 patients with LVSI and a DOI of ≤3 mm, one patient with squamous cell carcinoma had LNM (2.9%). Of the 57 patients with LVSI and a DOI of >3 mm and ≤5 mm, five patients had LNM (8.8%). There was no difference between these two groups (p=0.405).

No difference in risk of LNM was found between lesions seen by microscopy only and clinically visible lesions (3.8% versus 1.9% respectively, p=0.332).

1. Main results

Our results show an incidence of LNM of 2.6% in 389 patients with cervical cancer with ≤5 mm DOI and >7 mm diameter. The risk of LNM was increased in case of LVSI. Clinical visibility had no impact on the risk of LNM and in our study histological subtype was not associated with a higher risk of LNM. No cases of LNM were found in cervical cancers with a DOI ≤3 mm without LVSI. No difference in LNM was found between patients with LVSI and a DOI of ≤3 mm and patients with LVSI and a DOI of >3 mm and ≤5 mm.

2. Interpretations and implications

We could only identify three studies that analyzed the incidence of LNM in patients with 2009 stage IB cervical cancer with a DOI of ≤5 mm and >7 mm diameter (FIGO 2018 stage IA) [6, 7, 10]. However, these studies were done in smaller study populations than our study.

In a prospective cohort study on sentinel lymph node mapping, Sponholtz et al. [7] identified only 12 cases that would be classified as stage IA in the 2018 FIGO staging system of 233 patients with 2009 FIGO stage IB1. No LNM were found in this group of 12 patients. Stolnicu et al. [10] found an incidence of LNM of 0.8% in cervical tumors reallocated from stage IB (2009) to IA (2018). In a population-based study by Wenzel et al., [6] a 3.5% risk of LNM was found in a cohort of 170 patients.

Wenzel et al. [6] found a higher risk of LNM in adenocarcinoma (7.6%) than in squamous cell carcinoma (1.7%). In the absence of LVSI, they found no LNM in patients with squamous cell carcinoma. Consequently, they recommended omitting a lymph node dissection in patients with squamous cell carcinoma without LVSI. In contrast to these findings by Wenzel et al., [6] we observed no difference in risk of LNM between the histological types.

Many studies show that LVSI is a high-risk factor for LNM [11, 12, 13, 14, 15]. In the cohort from the study by Wenzel et al. [6] there was a higher incidence of LNM in tumors with LVSI (9.3%) compared to the group without LVSI (1.7%). We also found a higher risk of LNM in univariate analysis, confirming the importance of LVSI as a predictor for LNM. Both our study and the study by Wenzel et al. [6] found no LNM in cervical cancers without LVSI and with a DOI of ≤3 mm.

The official description of FIGO stage IA is “Invasive carcinoma that can be diagnosed only by microscopy, with maximum DOI ≤5 mm.” [1, 2]. This implies that clinically suspect lesions are always at least a stage IB. A study by Takahashi et al. [16] reclassified 25 FIGO 2009 stage IB tumors with a size of <2 cm. Of these 25 patients, 9 (36.0%) were reclassified as stage IA. Stolnicu et al. [10] performed a retrospective study on 416 cases with a FIGO 2009 stage IA carcinoma, and strictly used the FIGO guidelines by reclassifying all clinically visible tumors as stage IB. 30.3% of cases were reclassified as stage IA. These percentages are similar to our study (30.3%) but differ greatly from the percentage reclassified in the study by Sponholtz et al. [7] (5.2%). This may be due to the different interpretations of what “microscopy only” means and raises questions on how to define “clinically suspect.” For example, does “clinically suspect” also include lesions that are not suspect with the naked eye, but are suspect in colposcopy? How can we classify the patient where directed biopsies of a suspect lesion only show pre-invasive disease, but a conization shows a tumor with a DOI of 4 mm and a diameter of 20 mm? What about a cervical lesion which is not suspect at gynaecological examination, but is seen on MRI, and has a DOI of 3 mm and diameter of 8 mm?

In contrast to the study by Wenzel et al. [6] we also compared lesions which were clinically suspicious with lesions which could only be diagnosed by microscopy. Interestingly, our data did not show a difference in risk on LNM for the group of patients with clinically suspicious lesions versus the group that was diagnosed with microscopy only. This outcome challenges the condition “diagnosed by microscopy only” for stage IA in the FIGO classification.

Based on the results in our relatively large cohort, we recommend performing a pelvic lymph node assessment, by either a lymphadenectomy or a sentinel node procedure, in all patients with cervical cancer with a DOI of ≤5 mm and a diameter of >7 mm, except in the absence of LVSI and a DOI of ≤3 mm. Fig. 2 shows a flowchart with our recommendation on the performance of a pelvic lymph node assessment in this population. Importantly, a risk of LNM of 2.6% can be interpreted as a risk high enough to perform lymph node assessment but can also be seen as a relatively low risk. As lymph node assessment may result in morbidity, for example lymphedema, doctors and patients may take this into account in making treatment decisions regarding pelvic lymph node assessment in this group.

Fig. 2
Recommendations for lymph node assessment.
DOI, depth of invasion; LVSI, lymphovascular space invasion.

• Click for larger image
• Download as PowerPoint slide

Since there are only a few cases with LNM in our population, more research in larger and preferably prospective series is needed to allow firm conclusions on the treatment of cervical cancer with a DOI of ≤5 mm and a diameter of >7 mm.

3. Strengths and weaknesses

The strength of this research is the review of pathology specimen by expert pathologists in each medical center of cases with LNM and of cases in which the reports were inconclusive. Although all samples were initially or in retrospect examined by an expert pathologist, for this study we performed a revision of the lymph node positive cases. This allowed for the application of new insights and harmonization of methods. Especially in a retrospective study, a quality control like this can enhance the homogeneity of our results. The exclusion of cases and correction of data after revision demonstrates the importance of revision by expert pathologists. We could not find another study that collaborated so closely with expert pathologists to get more accurate results. Revision of all cases with positive lymph nodes resulted in a lower incidence of LNM and a higher incidence of LVSI in our research group. However, since we only revised cases with positive lymph nodes, we assume that the incidence of LNM could be higher after revision of all cases and exclusion of cases that do not fit the inclusion criteria.

Another strength is that it is a multicenter study which allowed for enhancement of generalizability and sharing of knowledge between these centers.

To our knowledge this is the largest study that evaluated the incidence of LNM in cervical cancers with a DOI of ≤5 mm and a diameter of >7 mm.

The limitation in this study is its retrospective nature. The quality of the data depends on the quality of the data in the medical records and pathology reports.

The incidence of LNM in patients with cervical cancer with ≤5 mm DOI and a diameter >7 mm is 2.6%. Clinical visibility is not associated with a higher risk of LNM. LVSI increases the risk of LNM. We recommend a pelvic lymph node assessment in all patients with cervical cancer with a DOI of ≤5 mm and a diameter of >7 mm, except in the absence of LVSI and a DOI of ≤3 mm. However, the incidence of LNM leaves room for shared decision-making regarding lymph node assessment.