We established a prognostic model to predict the long-term OS of patients with FIGO 2018 stage II-IIIC2r uterine cervical squamous cell carcinoma who received definitive radiotherapy. In this study, we found that factors such as the SCC-Ag level, pathologically poor differentiation, tumor volume derived from MRI prior to treatment, FIGO 2018 stage, and presence of chemotherapy were predictors of prognosis in these patients.
SCC-Ag levels are elevated in 28–88% of patients with uterine cervical squamous cell carcinoma (17). However, the cutoff value of SCC-Ag for predicting the prognosis of uterine cervical carcinoma remains controversial, with some researchers unable to demonstrate any predictive ability of the parameter at all (18). In contrast, several researchers demonstrated that SCC-Ag levels alone or in combination with other factors were significantly correlated with the prognosis and even had the capacity to predict the efficacy of treatment or risk of recurrence (19– 22). SCC-Ag, which was discovered by Kato et al., is a characteristic biomarker for squamous cell carcinoma (23). SCC-Ag expression emerges synchronously with the squamous formation of the uterine cervix and increases during the neoplastic transformation of the cervical squamous epithelium (24). Murakami et al. showed that SCC-Ag could promote radio-resistance of tumor cells by suppressing radiation-induced cell death (25).
Brambs et al. reexamined the histological slides of 467 patients with surgically treated FIGO stage IB1-IIB uterine cervical carcinoma and found that binary grading (grade 1/2 vs. grade 3) may be more suitable for evaluating prognostic survival than conventional tumor grading based on the degree of keratinization (26). This is also the reason we only considered poor differentiation, and not tumor grade, in our analysis, and our results are consistent with those of other studies. Studies by Xie et al. and Luo et al. found that patients with poor differentiation (grade 3) had a significantly worse OS than those with grade 1/2 uterine cervical carcinoma in the early stage (FIGO stages IA2-IIB) (27, 28). Using data from 31,536 women with uterine cervical squamous cell carcinoma extracted from the Surveillance, Epidemiology, and End Results (SEER) Program between 1983 and 2013, Matsuo et al. found that grade 3 tumors (poor differentiation) were independently associated with decreased cause-specific survival, especially among patients with stage II-III disease (29). These findings could be attributed to the keratin pattern being a component of aggregated cervical squamous cell carcinoma and being related to survival (30). On the contrary, Kumar et al. analyzed patients who were diagnosed with uterine cervical squamous cell carcinoma between 1988 and 2004, using limited data from the SEER Program, and figured out that nonkeratinized squamous cell carcinoma, rather than keratinized squamous cell carcinoma, may be more radiosensitive and associated with a better prognosis (31). Notably, the racial composition of the Asian population was 11.4%, and the proportion of poorly differentiated cases in our study was approximately 16% – a reason why the results of the above-mentioned studies are different from ours and why our findings are only applicable to Chinese patients.
FIGO 2018 is a clinical staging system based on physical exam and imaging. Gynecologists rely heavily on physical examination when evaluating primary tumors. However, palpation as a component of physical examination is a subjective method that can only determine the axial diameter of the tumor but cannot estimate the contribution of normal cervical tissue. Thus, clinical estimation of tumor size through palpation cannot adequately represent the actual tumor volume (32). Narayan et al. demonstrated in their study that tumor volume measured using MRI accurately reflected the extent of local disease and could be used as an objective measurement of the primary site of cervical cancer (33). Other researchers also demonstrated that an increase in tumor volume is associated with lymph node metastasis and poor prognosis (34, 35). Some investigators have even observed that MRI-derived tumor volume provides important prognostic information that is more accurate and useful than that provided by the FIGO staging system (36). However, the staging system used in all these studies was not the FIGO 2018 staging system. In our results, MRI-derived tumor volume was a critical prognostic factor for FIGO 2018 stage II-IIIC2r uterine cervical squamous cell carcinoma.
The FIGO staging system is widely used in the clinical management of uterine cervical carcinoma and is a paramount factor affecting the outcome of treatment. However, there are other prognostic factors that must be considered. Lymph node metastasis could strongly decrease the survival of patients with uterine cervical carcinoma, and in this regard, the FIGO 2018 staging system defined stage IIIC1 as pelvic lymph node metastasis and stage IIIC2 as para-aortic lymph node metastasis, both of which can be suffixed with the letter “r” or “p” to refer to a radiological or pathological finding, respectively (5). Therefore, we contrasted a nomogram using the FIGO 2018 staging system and other clinical factors, with emphasis on MRI-derived tumor volume. NRI and IDI are indices indicating how a model's predictive power improves after a new risk factor or factors are introduced. A value of > 0 indicates improvement. In this study, the NRI and IDI values for 3-year, 5-year, and 7-year OS were all > 0, suggesting that our model achieved a better predictive ability than the FIGO 2018 staging system. Thus, our nomogram could offer patients accurate individual predictions. Moreover, since chemotherapy as a therapeutic factor was included in our model, this means that chemotherapy may improve the OS of patients. This could be a very useful parameter to gynecologic oncologists when creating treatment plans and to patients when deciding to accept those plans. For instance, if a virtual 59-year-old patient with a pathologically confirmed uterine cervical squamous cell carcinoma, whose FIGO stage is Ⅲb, pretreatment SCC-Ag level is 15.0ng/mL, and MRI-derived tumor volume is 35 cm3, decides to undergo chemotherapy, the total score for all parameters calculated from the nomogram will be 139, and the predictive 3-year, 5-year, and 7-year OS will be 86%, 80%, and 78%, respectively. However, if she refuses to have chemotherapy, the total score will be 187, and the predictive 3-year, 5-year, and 7-year OS will be 73%, 63%, and 60%. Radiotherapy without chemotherapy is associated with an apparent decrease in OS.
There have been several nomograms established by other researchers to predict uterine cervical carcinoma following radiotherapy (9, 37, 38). Other researchers explored the predictive accuracy of the FIGO 2018 staging system and other significant prognostic factors; however, they have not investigated the value of MRI-derived tumor volume in predicting the prognosis of patients (39, 40). To our knowledge, this nomogram is the first long-term model to predict OS in patients with uterine cervical squamous cell carcinoma who received definitive radiotherapy, using the FIGO 2018 staging system and pretreatment tumor volume derived from MRI.
This study has some limitations. First, it is a retrospective study and is therefore prone to selection bias. Second, our prediction model was developed using data from a single institution and therefore needs to be further validated externally. Third, the chemotherapy regimens in this study were heterogeneous: Some patients, who had the opportunity to have surgery or had a strong willingness to undergo further treatment, received neoadjuvant chemotherapy. A proportion of patients received neoadjuvant chemotherapy due to unsatisfactory tumor shrinkage or abnormal SCC-Ag level after radiotherapy, with common iliac artery or para-aortic lymph node metastasis. Thus, we only analyzed the presence of chemotherapy as a factor in our research. Further stratified analysis of the specific chemotherapy regimens is necessary.