Skip to main content

ORIGINAL RESEARCH article

Front. Oncol., 16 August 2021
Sec. Gynecological Oncology
This article is part of the Research Topic Insights in Gynecological Oncology: 2021 View all 11 articles

Radiomic Score as a Potential Imaging Biomarker for Predicting Survival in Patients With Cervical Cancer

Handong Li&#x;Handong Li1†Miaochen Zhu&#x;Miaochen Zhu2†Lian Jian&#x;Lian Jian1†Feng BiFeng Bi1Xiaoye ZhangXiaoye Zhang2Chao FangChao Fang3Ying WangYing Wang2Jing Wang*Jing Wang4*Nayiyuan Wu*Nayiyuan Wu2*Xiaoping Yu*Xiaoping Yu1*
  • 1Department of Radiology, Hunan Cancer Hospital, The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, China
  • 2Central Laboratory, Hunan Cancer Hospital, The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, China
  • 3Department of Clinical Pharmaceutical Research Institution, Hunan Cancer Hospital, Affiliated Tumor Hospital of Xiangya Medical School of Central South University, Changsha, China
  • 4Gynecological Oncology Clinical Research Center, Hunan Cancer Hospital, Affiliated Tumor Hospital of Xiangya Medical School of Central South University, Changsha, China

Objectives: Accurate prediction of prognosis will help adjust or optimize the treatment of cervical cancer and benefit the patients. We aimed to investigate the incremental value of radiomics when added to the FIGO stage in predicting overall survival (OS) in patients with cervical cancer.

Methods: This retrospective study included 106 patients with cervical cancer (FIGO stage IB1–IVa) between October 2017 and May 2019. Patients were randomly divided into a training cohort (n = 74) and validation cohort (n = 32). All patients underwent contrast-enhanced computed tomography (CT) prior to treatment. The ITK-SNAP software was used to delineate the region of interest on pre-treatment standard-of-care CT scans. We extracted 792 two-dimensional radiomic features by the Analysis Kit (AK) software. Pearson correlation coefficient analysis and Relief were used to detect the most discriminatory features. The radiomic signature (i.e., Radscore) was constructed via Adaboost with Leave-one-out cross-validation. Prognostic models were built by Cox regression model using Akaike information criterion (AIC) as the stopping rule. A nomogram was established to individually predict the OS of patients. Patients were then stratified into high- and low-risk groups according to the Youden index. Kaplan–Meier curves were used to compare the survival difference between the high- and low-risk groups.

Results: Six textural features were identified, including one gray-level co-occurrence matrix feature and five gray-level run-length matrix features. Only the FIGO stage and Radscore were independent risk factors associated with OS (p < 0.05). The C-index of the FIGO stage in the training and validation cohorts was 0.703 (95% CI: 0.572–0.834) and 0.700 (95% CI: 0.526–0.874), respectively. Correspondingly, the C-index of Radscore was 0.794 (95% CI: 0.707–0.880) and 0.754 (95% CI: 0.623–0.885). The incorporation of the FIGO stage and Radscore achieved better performance, with a C-index of 0.830 (95% CI: 0.738–0.922) and 0.772 (95% CI: 0.615–0.929), respectively. The nomogram based on the FIGO stage and Radscore could individually predict the OS probability with good discrimination and calibration. The high-risk patients had shorter OS compared with the low-risk patients (p < 0.05).

Conclusion: Radiomics has the potential for noninvasive risk stratification and may improve the prediction of OS in patients with cervical cancer when added to the FIGO stage.

Introduction

Cervical cancer is one of the fourth most common female malignancies worldwide (1). More than 80% of patients are typically diagnosed at a locally advanced stage (2). Five-year overall survival (OS) can be significantly distinct, ranging from 80% (stage IB) to 15% (stage IVa-b) (3). Despite the fact that outcomes of cervical cancer had been improved with multimodality treatment, around 30%–40% of patients still suffer from recurrence (4). Thus, it is of great significance to identify high-risk patients who may benefit from aggressive treatment.

The International Federation of Gynecology and Obstetrics (FIGO) stage has been established as the most crucial prognostic factor for cervical cancer (5). The treatment modality choice is mainly based on the FIGO stage and N staging (6). However, clinical outcomes are markedly different among patients with similar stages (6). Imaging plays an essential role in the pre-treatment evaluation of cervical cancer. However, conventional medical images only provide structural information of cancer; it fails to detect the intratumoral heterogeneity associated with treatment response and prognosis (7). Thus, the search for new non-invasive biomarkers with the potential to offer more specific tumor characterization before therapy is urgently needed, which may inform clinicians to make a more individualized treatment plan.

Statistical models, medical images, and machine learning have been widely used for outcome prediction in cervical cancer. Machine learning has merits in dealing with the complexity of high-dimensional data and discovering prognostic factors. Radiomics refers to a variety of mathematical methods such as machine learning that converts digital medical images into a huge number of minable high-dimensional features for cancer diagnosis or prediction (8, 9). Radiomic signature can be used as a surrogate biomarker for biological tumor traits such as tumor morphology and intratumor heterogeneity (10, 11). Currently, radiomics has been used to predict tumor stage, histological type, lymph node metastasis, relapse, and survival in patients with cervical cancer (12). However, the additional value of radiomics to the FIGO stage in prognostication of cervical cancer remains unclear. Thus, this study aimed to develop and validate a radiomic model for predicting survival in patients with cervical cancer.

Materials and Methods

Patient Cohort

This retrospective study included patients with a diagnosis of cervical cancer between October 2017 and May 2019 at our institution. Inclusion criteria were (1) patients with histologically confirmed cervical cancer, (2) patients with tumor staged IB1-IVa (FIGO 2009 definition), (3) patients who were not previously treated with any anti-cancer treatment, and (4) patients who underwent a pre-treatment contrast-enhanced computed tomography (CT) scan. Exclusion criteria were (1) patients with a history of previous chemotherapy or radiotherapy, (2) patients with a diagnosis of other cancers, or (3) patients with distant metastatic disease (para-aortic nodes involvement was not included). Institutional ethics review board approval was acquired for this study and written informed consent was waived. Finally, a total of 106 patients (mean age, 63.8 years) were included in this study. Eligible patients for the radiomic study were randomly divided into a training cohort (n = 74) and a validation cohort (n = 32). The clinical information of patients was collected from electronic medical records, including age, FIGO stage, histological type, differentiation, lymph node metastasis (LNM), and treatment regimens. Of note, a lymph node with a short-axis diameter larger than 10 mm was considered to be metastasis (13).

Treatment Characteristics and Follow-Up

All patients were treated with image-guided external beam radiotherapy (EBRT) and brachytherapy (BT), to a total dose of 85–90 Gy (EQD2, equivalent dose in 2 Gy single dose fractions). EBRT was delivered in 1.8–2.0 Gy/fraction, to a range of 45–50 Gy, using a 3D conformal technique. BT boost was volumetrically planned and delivered as weekly high-dose-rate fractions of 8 Gy EQD2 each after 15 times of EBRT. The external irradiation was not performed on the day of intracavitary and interstitial after loading BT. The radiation volumes covered the pelvic cavity and, if clinically indicated, the para-aortic and/or inguinal nodal regions. CT-positive lymph nodes were simultaneously boosted to a total dose of 55–60 Gy. Concurrent weekly chemotherapy with cisplatin (40 mg/m2) was delivered for 4–6 weeks or carboplatin (AUC = 2) when feasible. The endpoint OS was defined as the interval from the date of treatment to death from any cause. The patients were followed up until November 19, 2020.

CT Image Acquisitions

CT scans were performed on a 64-row CT scanner (Somatom definition AS large-aperture, Siemens Healthcare) using the following parameters: 120-kVp tube voltage, 252-mAs tube current, a field of view (FOV) of 384 × 384 mm2, a width of detector of 40 mm, a beam pitch of 0.6, a gantry rotation time of 0.5 s, and a slice thickness of 5 mm. Iohexol (350 mg I/ml) was administrated with a rate of 2.5–3.0 ml/s through the elbow vein by a high-pressure injector. Enhanced CT images were obtained at 30 s after injection.

Radiomic Analysis

Our radiomic workflow is illustrated in Figure 1.

FIGURE 1
www.frontiersin.org

Figure 1 Schematic diagram exhibition of the radiomic workflow. A radiomic study design and workflow mainly include (I) Image segmentation, (II) Radiomic feature extraction, (III) Dimension reduction and feature selection, (IV) Statistics analysis and model building. CT, computed tomography; GLCM, gray-level co-occurrence matrix; GLZSM, gray-level size zone matrix; GLRLM, gray-level run-length matrix; PCC, Pearson correction coefficient.

Image Segmentation

We used an open-source ITK-SNAP software (version 3.6.0, www.itksnap.org) for manual segmentation on CT images (Figure 2). Tumor lesion segmentation was performed on the maximum level of tumor by a radiologist (with 6-year experience) and subsequently reviewed by a board-certified radiologist (>10 years experience).

FIGURE 2
www.frontiersin.org

Figure 2 Illustration of tumor segmentation on the maximum level of tumor. (A) Raw pre-contrast image; (B) pre-contrast image after delineation; (C) raw post-contrast image; and (D) post-contrast image after delineation.

Feature Extraction

The radiomic features were automatically calculated by AK software (Artificial Intelligence Kit, GE Life Sciences, AA R&D team, Shanghai, China), which comply with the standards set by the Image Biomarker Standardization Initiative. In total, 792 radiomic features were extracted from pre-contrast and post-contrast CT images, including (1) histogram features, such as energy, entropy, uniformity, skewness, and kurtosis; (2) form factor features, such as sphericity, surface area, compactness, surface volume ratio, maximum 3D diameter, spherical disproportion, volume CC, and volume MM; and (3) texture features including gray-level co-occurrence matrix (GLCM), gray-level size zone matrix (GLZSM), gray-level run-length matrix (GLRLM), and Haralick parameters. The offsets of GLCM and GLRLM were 1, 4, and 7.

Feature Pre-Processing and Selection

Feature pre-processing was done in two steps: step 1—outliers and null values were replaced by mean values, and step 2—values standardization was carried out to eliminate the influence of the dimension (14). Feature selection is a critically important step for better generalization of models because high-dimensional data usually comprise a large number of irrelevant, redundant, and noisy features, which may result in the curse of dimensionality and model overfitting (15). In terms of feature selection, Pearson correction coefficient (PCC) analysis was used to assess the correlation between feature pairs, and one feature was randomly excluded from each pair with a correlation coefficient >0.9. After this process, the dimension of the variable space was reduced, and each variable was independent of each other. Then, we used Relief, a feature weighting algorithm, to detect the most discriminatory features. Finally, radiomic signature (i.e., Radscore) based on the selected features was constructed via Adaboost with Leave-one-out cross-validation.

Prognostic Model Construction

Variables with missing data of more than 20% were excluded for analysis. The significant variables were identified by the univariate and multivariate Cox regression model. The variables with p < 0.10 in the univariate analysis were entered into multivariate analysis to obtain independent risk factors for OS (p < 0.05). The prognostic models were constructed by Cox proportional hazard model, using the Akaike information criterion (AIC) as the stopping rule. For convenient use by clinicians, we built a nomogram to individually predict the OS of patients based on the optimal prognostic model. The calibration curves reflecting the goodness of fit of the nomograms generated were assessed by plotting the predicted probabilities against the observed event proportions. The receiver operating characteristic (ROC) curve was used to determine cutoff values of models according to the Youden index to generate Kaplan–Meier curves for OS in the training and validation cohorts. The Log-rank test was used for comparisons in the Kaplan–Meier curves.

Statistical Analysis

The clinicopathologic characteristics were assessed by applying two-sample t-test, chi-squared test, or Mann–Whitney U-test in the training and validation cohorts, where appropriate. The discrimination of each model was quantified by the C-index and 95% confidence interval (CI). The models were subjected to bootstrapping validation (1000 bootstrap resamples) to obtain a relatively corrected C-index. All statistical analyses were performed using 3.6.0 R software (http://www.Rproject.org). A two-tailed p < 0.05 was considered statistically significant.

Results

The Patient Cohort of the Radiomic Study

There were no significant differences in the clinicopathologic characteristics between the training and validation cohorts (all P values >0.05) (Table 1).

TABLE 1
www.frontiersin.org

Table 1 Patients’ characteristics.

Feature Selection and Radscore Construction

A total of 428 radiomic features were retained after PCC analysis; only six texture features were then selected by Relief, including one GLCM feature (pre_GLCMEntropy_AllDirection_offset7_SD) and five GLRLM features (post_LongRunEmphasis_angle135_offset1.1, pre_LongRunEmphasis_AllDirection_offset1_SD.1, pre_LongRunLowGreyLevelEmphasis_AllDirection_offset1, post_LongRunEmphasis_angle45_offset1, and post_LongRunHighGreyLevelEmphasis_AllDirection_offset4_SD.1). Pre_ and post_ denote the pre-contrast and post-contrast features, respectively. The histogram and form factor features were omitted. The Radscore was built by the Adaboost classifier based on the six textural features.

Independent Indicators of OS

Multivariate Cox regression analysis showed that among clinicopathologic characteristics, only the FIGO stage was an independent factor of OS (p < 0.002). The hazard ratio (HR) for the FIGO stage was 3.24 (95% CI: 1.55–6.76). When adding the Radscore to the aforementioned features, only Radscore was an independent indicator of OS (p = 0.002). The HRs for the FIGO stage and Radscore were 2.07 (95% CI: 0.89–4.83) and 153.0 (95% CI: 6.29–3721.16), respectively.

Performance of Prognostic Models

The C-index of the FIGO stage in the training and validation cohorts was 0.703 (95% CI: 0.572–0.834) and 0.700 (95% CI: 0.526–0.874), respectively (Table 2). The C-index of Radscore was 0.794 (95% CI: 0.707–0.880) and 0.754 (95% CI: 0.623–0.885), respectively (Table 2). The incorporation of the FIGO stage and Radscore achieved better performance, with a C-index of 0.830 (95% CI: 0.738–0.922) and 0.772 (95% CI: 0.615–0.929), respectively (Table 2). The nomogram based on the FIGO stage and Radscore could individually predict the OS probability (Figure 3). Calibration curves demonstrated good agreement between the predicted probability and observed probability (Figure 3).

TABLE 2
www.frontiersin.org

Table 2 The performance of FIGO stage, Radscore, and the combined model for OS evaluation in patients with cervical cancer.

FIGURE 3
www.frontiersin.org

Figure 3 The nomogram (A) based on the FIGO stage and Radscore was used to estimate OS individually, along with the assessment of the model calibration. Calibration curves for the nomogram to the 1-year and 2-year OS rate in the training (B) and validation cohorts (C). FIGO, International Federation of Gynecology and Obstetrics; OS, overall survival.

Kaplan–Meier Survival Analysis

Kaplan–Meier curves demonstrate that by using a cutoff value of 0.59, Radscore alone could stratify patients into the low- and high-risk groups, with significant OS differences. The high-risk patients had significantly shorter OS than the low-risk patients (training cohort: p < 0.0001; validation cohort: p = 0.03). The combined FIGO stage and Radscore could achieve better risk stratification using a cutoff value of 0.57 (training cohort: p < 0.0001, validation cohort: p = 0.02) (Figure 4).

FIGURE 4
www.frontiersin.org

Figure 4 Kaplan–Meier curves of the high- and low-risk patients stratified by the combined model in the training cohort and validation cohort.

Discussion

In this radiomic study, we demonstrated that capturing intratumoral radiomics from pretherapy CT images could significantly improve the performance of the FIGO stage for OS evaluation. The combined model yielded the optimal performance for OS prediction in patients with cervical cancer. We established a radiomic nomogram derived from a combined model to individually estimate the OS of patients. The nomogram could successfully stratify these patients into high-risk and low-risk subgroups, and these two subgroups had different OS in both training and validation cohorts.

Prognosis evaluation in advance will greatly improve the treatment outcome in patients with cervical cancer. The most relevant tumor-related prognostic factors for locally advanced cervical cancer are tumor size at diagnosis, FIGO stage, and LNM (6, 16). However, our work showed that only the FIGO stage was an independent risk factor of OS in stage IB1–IVa cervical cancer. FIGO stage is the current standard staging system for cervical cancer based on findings from physical inspection and imaging examinations (17). The limitations of the FIGO stage are that it depends on many examinations and its classification is not objective enough (18). Therefore, an objective and efficient evaluation method that complements cervical cancer staging is of clinical significance.

The current radiomic workflow consists of adding quantitative data to visual analysis rather than replacing it entirely (19). In the field of oncology, translational approaches are increasingly explored, for instance, Radiogenomics. Such studies usually need “big data” approaches and collaborative research (20). In this work, we proposed a new approach, translational but able to be quickly implemented for clinicians. To the best of our knowledge, this is the first study to integrate accessible FIGO stage and radiomic features extracted from pre-treatment CT images to predict the survival of cervical cancer. Radscore may be a more powerful predictor compared with the FIGO stage. With the help of Radscore, the prognostic value of the FIGO stage was significantly improved, leading to better risk stratification of patients. This score may be used for adaptive strategies to improve patient outcomes.

Radiomics has been involved in many aspects of cervical cancer, including prediction of tumor staging (21, 22), histological grading (23, 24), lymphovascular space invasion (LVI) (25), LNM (2636), treatment response (3742), and outcome (4348) based on multimodal imaging tools (e.g., CT, MRI, and PET/CT). In clinical practice, the issues of most concern may be the evaluation of LNM, treatment response, and survival prediction. Hence, most radiomic studies focused on the evaluation of these in a non-invasive way. LNM is a key prognostic factor that affects the treatment decision and survival in cervical cancer. Song et al. combined the Radscore and morphological features of lymph nodes to assess the LNM in cervical cancer patients (26). Some studies developed radiomic models based on Radscore and clinical factors (e.g., MRI-reported lymph node status and FIGO stage) for predicting LNM in cervical cancer (2736). Treatment response to neoadjuvant or concurrent chemotherapy may have a significant impact on patient management by identifying tailored approaches for patient subgroups to achieve a better clinical outcome. All relevant studies showed the additional value of radiomics to clinical factors and its potential to screen patients who are sensitive to chemotherapy (3742). Survival evaluation is an eternal proposition for the field of oncology, which may benefit treatment strategies and follow-up plans. Ferreira et al. tested the feasibility of PET radiomic features combined with clinical information in predicting disease-free survival (DFS) in patients with cervical cancer (43). Fang et al. used MRI-based Radscore, LNM, and LVI to predict the DFS of early-stage cervical cancer (45). Lucia et al. found that the combination of PET/CT and MRI could result in more favorable survival prediction (47). Another study suggested that Radscore could enhance the prediction efficiency of conventional PET parameters in cervical cancer (48). In this current study, the results indicated that Radscore was an independent risk factor strongly associated with OS of cervical cancer and only texture features contributed to the survival prediction in cervical cancer. The HR of the Radscore was significantly higher than that of the FIGO stage, suggesting that the Radscore may provide more prognostic information than the FIGO stage in predicting the OS.

This study also had some limitations. First, selection bias was inevitable because of a retrospective study. Second, the sample size was small for a radiomic study. However, we performed a bootstrap to obtain corrected results. Third, the overall treatment times were not readily available, which is a well-recognized factor in cervical cancer outcomes. Fourth, this study had no local/regional failures, recognizing the short follow-up time. The additional evaluation of local control could be more clinically significant. Fifth, our radiomic features were extracted from CT images; those features from MRI or PET/CT images may provide different information about intratumoral heterogeneity. Sixth, the image analysis was based on the maximum level of the tumor instead of the whole tumor. However, some previous studies (45, 49) showed that the predictive performance of features extracted from the maximum level of the tumor was higher than that of those features extracted from the whole tumor. Two-dimensional features may increase the robustness of features compared with three-dimensional features. Seventh, the biological interpretation of radiomics remains an open question warranting further investigation (50). Finally, this is a single-institution study that needs external validation of the findings.

In conclusion, this present study provided a combined model, incorporating the FIGO stage and a Radscore derived from CT-based textural features, with favorable performance, and the study developed a noninvasive radiomic nomogram based on the results of the combined model for the pretherapy and personalized estimation of OS in patients with cervical cancer. In addition, the combined model serves as a collection of potential biomarkers and perfectly stratifies these patients into high-risk and low-risk subgroups. The identification of high-risk patients at diagnosis can allow tailored treatments involving higher doses of radiation boost, consolidation chemotherapy, and/or adjuvant hysterectomy, when indicated, and should be confirmed in external cohorts and prospective studies.

Data Availability Statement

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.

Ethics Statement

The studies involving human participants were reviewed and approved by Hunan Cancer Hospital, The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University. Written informed consent was waived.

Author Contributions

HL, MZ, and LJ designed the study. HL, MZ, and LJ contributed to the conception of the study and completed the manuscript together. JL, FB, and XZ contributed significantly to statistical analysis. YW and JW completed the following-up information. LJ, NW, and XY drafted the manuscript. All authors contributed to the article and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin (2021) 70(4):313. doi: 10.3322/caac.21660

CrossRef Full Text | Google Scholar

2. Cohen PA, Jhingran A, Oaknin A, Denny L. Cervical Cancer. Lancet (2019) 393:169–82. doi: 10.1016/S0140-6736(18)32470-X

PubMed Abstract | CrossRef Full Text | Google Scholar

3. Edge SB, Compton CC. The American Joint Committee on Cancer: The 7th Edition of the AJCC Cancer Staging Manual and the Future of TNM. Ann Surg Oncol (2010) 17:1471–4. doi: 10.1245/s10434-010-0985-4

PubMed Abstract | CrossRef Full Text | Google Scholar

4. Herrera FG, Breuneval T, Prior JO, Bourhis J, Ozsahin M. [18f]FDG-PET/CT Metabolic Parameters as Useful Prognostic Factors in Cervical Cancer Patients Treated With Chemo-Radiotherapy. Radiat Oncol (2016) 11(1):43. doi: 10.1186/s13014-016-0614-x

PubMed Abstract | CrossRef Full Text | Google Scholar

5. Monk BJ, Tewari KS, Koh WJ. Multimodality Therapy for Locally Advanced Cervical Carcinoma: State of the Art and Future Directions. J Clin Oncol (2007) 25:2952–65. doi: 10.1200/JCO.2007.10.8324

PubMed Abstract | CrossRef Full Text | Google Scholar

6. Rose PG, Java J, Whitney CW, Stehman FB, Lanciano R, Thomas GM, et al. Nomograms Predicting Progression-Free Survival, Overall Survival, and Pelvic Recurrence in Locally Advanced Cervical Cancer Developed From an Analysis of Identifiable Prognostic Factors in Patients From NRG Oncology/Gynecologic Oncology Group Randomized Trials of Chemoradiotherapy. J Clin Oncol (2015) 33(19):2136. doi: 10.1200/JCO.2014.57.7122

PubMed Abstract | CrossRef Full Text | Google Scholar

7. Caudell JJ, Torres-Roca JF, Gillies RJ, Enderling H, Kim S, Rishi A, et al. The Future of Personalised Radiotherapy for Head and Neck Cancer. Lancet Oncol (2017) 18:e266–73. doi: 10.1016/S1470-2045(17)30252-8

PubMed Abstract | CrossRef Full Text | Google Scholar

8. Bi WL, Hosny A, Schabath MB, Giger ML, Birkbak NJ, Mehrtash A, et al. Artificial Intelligence in Cancer Imaging: Clinical Challenges and Applications. CA Cancer J Clin (2019) 69:127–57. doi: 10.3322/caac.21552

PubMed Abstract | CrossRef Full Text | Google Scholar

9. Lambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, et al. Radiomics: The Bridge Between Medical Imaging and Personalized Medicine. Nat Rev Clin Oncol (2017) 14:749–62. doi: 10.1038/nrclinonc.2017.141

PubMed Abstract | CrossRef Full Text | Google Scholar

10. Aerts HJ, Velazquez ER, Leijenaar RT, Parmar C, Grossmann P, Carvalho S, et al. Decoding Tumour Phenotype by Noninvasive Imaging Using a Quantitative Radiomics Approach. Nat Commun (2014) 5:4006. doi: 10.1038/ncomms5644

PubMed Abstract | CrossRef Full Text | Google Scholar

11. Lambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, van Stiphout RG, Granton P, et al. Radiomics: Extracting More Information From Medical Images Using Advanced Feature Analysis. Eur J Cancer (2012) 48(4):441–6. doi: 10.1016/j.ejca.2011.11.036

PubMed Abstract | CrossRef Full Text | Google Scholar

12. Ai Y, Zhu H, Xie C, Jin X. Radiomics in Cervical Cancer: Current Applications and Future Potential. Crit Rev Oncol Hematol (2020) 152:102985. doi: 10.1016/j.critrevonc.2020.102985

PubMed Abstract | CrossRef Full Text | Google Scholar

13. McMahon CJ, Rofsky NM, Pedrosa I. Lymphatic Metastases From Pelvic Tumors: Anatomic Classification, Characterization, and Staging. Radiology (2010) 254:31–46. doi: 10.1148/radiol.2541090361

PubMed Abstract | CrossRef Full Text | Google Scholar

14. Zhang Y, Yu S, Zhang L, Kang L. Radiomics Based on CECT in Differentiating Kimura Disease From Lymph Node Metastases in Head and Neck: A Non-Invasive and Reliable Method. Front Oncol (2020) 10:1121. doi: 10.3389/fonc.2020.01121

PubMed Abstract | CrossRef Full Text | Google Scholar

15. Limkin EJ, Sun R, Dercle L, Zacharaki EI, Robert C, Reuzé S, et al. Promises and Challenges for the Implementation of Computational Medical Imaging (Radiomics) in Oncology. Ann Oncol (2017) 28:1191–206. doi: 10.1093/annonc/mdx034

PubMed Abstract | CrossRef Full Text | Google Scholar

16. Kristensen GB, Abeler VM, Risberg B, Tropé C, Bryne M. Tumor Size, Depth of Invasion, and Grading of the Invasive Tumor Front are the Main Prognostic Factors in Early Squamous Cell Cervical Carcinoma. Gynecoloncol (1999) 74(2):245–51. doi: 10.1006/gyno.1999.5420

CrossRef Full Text | Google Scholar

17. Creasman WT. New Gynecologic Cancer Staging. Obstetr Gynecol (1990) 75(2):287–8.

Google Scholar

18. Nordström B, Strang P, Lindgren A, Bergström R, Tribukait B. Carcinoma of the Endometrium: Do the Nuclear Grade and DNA Ploidy Provide More Prognostic Information Than do the FIGO and WHO Classifications? Int J Gynecol Pathol (1996) 15(3):191–201. doi: 10.1097/00004347-199607000-00002

PubMed Abstract | CrossRef Full Text | Google Scholar

19. Aerts HJWL. The Potential of Radiomic-Based Phenotyping in Precision Medicine: A Review. JAMA Oncol (2016) 2(12):1636–42. doi: 10.1001/jamaoncol.2016.2631

PubMed Abstract | CrossRef Full Text | Google Scholar

20. Andreassen CN, Schack LMH, Laursen LV, Alsner J. Radiogenomics – Current Status, Challenges and Future Directions. Cancer Lett (2016) 382:127–36. doi: 10.1016/j.canlet.2016.01.035

PubMed Abstract | CrossRef Full Text | Google Scholar

21. Mu W, Chen Z, Liang Y, Shen W, Yang F, Dai R, et al. Staging of Cervical Cancer Based on Tumor Heterogeneity Characterized by Texture Features on 18F-FDG PET Images. Phys Med Biol (2015) 60(13):5123. doi: 10.1088/0031-9155/60/13/5123

PubMed Abstract | CrossRef Full Text | Google Scholar

22. Umutlu L, Nensa F, Demircioglu A, Antoch G, Herrmann K, Forsting M, et al. Radiomics Analysis of Multiparametric PET/MRI for N- and M-Staging in Patients With Primary Cervical Cancer. Rofo (2020) 192:754–63. doi: 10.1055/a-1100-0127

PubMed Abstract | CrossRef Full Text | Google Scholar

23. Liu Y, Zhang Y, Cheng R, Liu S, Qu F, Yin X, et al. Radiomics Analysis of Apparent Diffusion Coefficient in Cervical Cancer: A Preliminary Study on Histological Grade Evaluation. J Magn Reson Imaging (2019) 49:280–90. doi: 10.1002/jmri.26192

PubMed Abstract | CrossRef Full Text | Google Scholar

24. Shen WC, Chen SW, Liang JA, Hsieh TC, Yen KY, Kao CH. [18] Fluorodeoxyglucose Positron Emission Tomography for the Textural Features of Cervical Cancer Associated With Lymph Node Metastasis and Histological Type. Eur J Nucl Med Mol I (2017) 44(10):1721–31. doi: 10.1007/s00259-017-3697-1

CrossRef Full Text | Google Scholar

25. Jiang X, Li J, Kan Y, Yu T, Chang S, Sha X, et al. MRI Based Radiomics Approach With Deep Learning for Prediction of Vessel Invasion in Early-Stage Cervical Cancer. IEEE/ACM Trans Comput Biol Bioinform (2020) 18(3):995–1002. doi: 10.1109/TCBB.2019.2963867

CrossRef Full Text | Google Scholar

26. Song J, Hu Q, Ma Z, Zhao M, Chen T, Shi H. Feasibility of TWI-MRI-Based Radiomics Nomogram for Predicting Normal-Sized Pelvic Lymph Node Metastasis in Cervical Cancer Patients. Eur Radiol (2021). doi: 10.1007/s00330-021-07735-x

CrossRef Full Text | Google Scholar

27. Deng X, Liu M, Sun J, Li M, Liu D, Li L, et al. Feasibility of MRI-Based Radiomics Features for Predicting Lymph Node Metastases and VEGF Expression in Cervical Cancer. Eur J Radiol (2021) 134:109429. doi: 10.1016/j.ejrad.2020.109429

PubMed Abstract | CrossRef Full Text | Google Scholar

28. Chen X, Liu W, Thai TC, Castellano T, Gunderson CC, Moore K, et al. Developing a New Radiomics-Based CT Image Marker to Detect Lymph Node Metastasis Among Cervical Cancer Patients. Comput Methods Progr BioMed (2020) 197:105759. doi: 10.1016/j.cmpb.2020.105759

CrossRef Full Text | Google Scholar

29. Hou L, Zhou W, Ren J, Du X, Xin L, Zhao X, et al. Radiomics Analysis of Multiparametric MRI for the Preoperative Prediction of Lymph Node Metastasis in Cervical Cancer. Front Oncol (2020) 10:1393. doi: 10.3389/fonc.2020.01393

PubMed Abstract | CrossRef Full Text | Google Scholar

30. Dong T, Yang C, Cui B, Zhang T, Sun X, Song K, et al. Development and Validation of a Deep Learning Radiomics Model Predicting Lymph Node Status in Operable Cervical Cancer. Front Oncol (2020) 10:464. doi: 10.3389/fonc.2020.00464

PubMed Abstract | CrossRef Full Text | Google Scholar

31. Xiao M, Ma F, Li Y, Li Y, Li M, Zhang G, et al. Multiparametric MRI-Based Radiomics Nomogram for Predicting Lymph Node Metastasis in Early-Stage Cervical Cancer. J Magn Reson Imaging (2020) 52:885–96. doi: 10.1002/jmri.27101

PubMed Abstract | CrossRef Full Text | Google Scholar

32. Jin X, Ai Y, Zhang J, Zhu H, Jin J, Teng Y, et al. Noninvasive Prediction of Lymph Node Status for Patients With Early-Stage Cervical Cancer Based on Radiomics Features From Ultrasound Images. Eur Radiol (2020) 30:4117–24. doi: 10.1007/s00330-020-06692-1

PubMed Abstract | CrossRef Full Text | Google Scholar

33. Wu Q, Wang S, Chen X, Wang Y, Dong L, Liu Z, et al. Radiomics Analysis of Magnetic Resonance Imaging Improves Diagnostic Performance of Lymph Node Metastasis in Patients With Cervical Cancer. Radiother Oncol (2019) 138:141–8. doi: 10.1016/j.radonc.2019.04.035

PubMed Abstract | CrossRef Full Text | Google Scholar

34. Wang T, Gao T, Yang J, Yan X, Wang Y, Zhou X, et al. Preoperative Prediction of Pelvic Lymph Nodes Metastasis in Early-Stage Cervical Cancer Using Radiomics Nomogram Developed Based on T2-Weighted MRI and Diffusion-Weighted Imaging. Eur J Radiol (2019) 114:128–35. doi: 10.1016/j.ejrad.2019.01.003

PubMed Abstract | CrossRef Full Text | Google Scholar

35. Li Z, Li H, Wang S, Dong D, Yin F, Chen A, et al. MR-Based Radiomics Nomogram of Cervical Cancer in Prediction of the Lymph-Vascular Space Invasion Preoperatively. J Magn Reson Imaging (2019) 49:1420–6. doi: 10.1002/jmri.26531

PubMed Abstract | CrossRef Full Text | Google Scholar

36. Kan Y, Dong D, Zhang Y, Jiang W, Zhao N, Han L, et al. Radiomic Signature as a Predictive Factor for Lymph Node Metastasis in Early-Stage Cervical Cancer. J Magn Reson Imaging (2019) 49:304–10. doi: 10.1002/jmri.26209

PubMed Abstract | CrossRef Full Text | Google Scholar

37. Gui B, Autorino R, Miccò M, Nardangeli A, Pesce A, Lenkowicz J, et al. Pretreatment MRI Radiomics Based Response Prediction Model in Locally Advanced Cervical Cancer. Diagnost (Basel) (2021) 11(4). doi: 10.3390/diagnostics11040631

CrossRef Full Text | Google Scholar

38. Liu D, Zhang X, Zheng T, Shi Q, Cui Y, Wang Y, et al. Optimisation and Evaluation of the Random Forest Model in the Efficacy Prediction of Chemoradiotherapy for Advanced Cervical Cancer Based on Radiomics Signature From High-Resolution T2 Weighted Images. Arch Gynecol Obstet (2021) 303:811–20. doi: 10.1007/s00404-020-05908-5

PubMed Abstract | CrossRef Full Text | Google Scholar

39. Fang M, Kan Y, Dong D, Yu T, Zhao N, Jiang W, et al. Multi-Habitat Based Radiomics for the Prediction of Treatment Response to Concurrent Chemotherapy and Radiation Therapy in Locally Advanced Cervical Cancer. Front Oncol (2020) 10:563. doi: 10.3389/fonc.2020.00563

PubMed Abstract | CrossRef Full Text | Google Scholar

40. Tian X, Sun C, Liu Z, Li W, Duan H, Wang L, et al. Prediction of Response to Preoperative Neoadjuvant Chemotherapy in Locally Advanced Cervical Cancer Using Multicenter CT-Based Radiomic Analysis. Front Oncol (2020) 10:77. doi: 10.3389/fonc.2020.00077

PubMed Abstract | CrossRef Full Text | Google Scholar

41. Sun C, Tian X, Liu Z, Li W, Li P, Chen J, et al. Radiomic Analysis for Pretreatment Prediction of Response to Neoadjuvant Chemotherapy in Locally Advanced Cervical Cancer: A Multicentre Study. EBioMedicine (2019) 46:160–9. doi: 10.1016/j.ebiom.2019.07.049

PubMed Abstract | CrossRef Full Text | Google Scholar

42. Bowen SR, Yuh WTC, Hippe DS, Wu W, Partridge SC, Elias S, et al. Tumor Radiomic Heterogeneity: Multiparametric Functional Imaging to Characterize Variability and Predict Response Following Cervical Cancer Radiation Therapy. J Magn Reson Imaging (2018) 47:1388–96. doi: 10.1002/jmri.25874

PubMed Abstract | CrossRef Full Text | Google Scholar

43. Ferreira M, Lovinfosse P, Hermesse J, Decuypere M, Rousseau C, Lucia F, et al. [F]FDG PET Radiomics to Predict Disease-Free Survival in Cervical Cancer: A Multi-Scanner/Center Study With External Validation. Eur J Nucl Med Mol Imaging (2021). doi: 10.1007/s00259-021-05303-5

CrossRef Full Text | Google Scholar

44. Park SH, Hahm MH, Bae BK, Chong GO, Jeong SY, Na S, et al. Magnetic Resonance Imaging Features of Tumor and Lymph Node to Predict Clinical Outcome in Node-Positive Cervical Cancer: A Retrospective Analysis. Radiat Oncol (2020) 15:86. doi: 10.1186/s13014-020-01502-w

PubMed Abstract | CrossRef Full Text | Google Scholar

45. Fang J, Zhang B, Wang S, Jin Y, Wang F, Ding Y, et al. Association of MRI-Derived Radiomic Biomarker With Disease-Free Survival in Patients With Early-Stage Cervical Cancer. Theranostics (2020) 10:2284–92. doi: 10.7150/thno.37429

PubMed Abstract | CrossRef Full Text | Google Scholar

46. Lucia F, Visvikis D, Vallières M, Desseroit MC, Miranda O, Robin P, et al. External Validation of a Combined PET and MRI Radiomics Model for Prediction of Recurrence in Cervical Cancer Patients Treated With Chemoradiotherapy. Eur J Nucl Med Mol Imaging (2019) 46:864–77. doi: 10.1007/s00259-018-4231-9

PubMed Abstract | CrossRef Full Text | Google Scholar

47. Lucia F, Visvikis D, Desseroit MC, Miranda O, Malhaire JP, Robin P, et al. Prediction of Outcome Using Pretreatment F-FDG PET/CT and MRI Radiomics in Locally Advanced Cervical Cancer Treated With Chemoradiotherapy. Eur J Nucl Med Mol Imaging (2018) 45:768–86. doi: 10.1007/s00259-017-3898-7

PubMed Abstract | CrossRef Full Text | Google Scholar

48. Schernberg A, Reuze S, Orlhac F, Buvat I, Dercle L, Sun R, et al. A Score Combining Baseline Neutrophilia and Primary Tumor SUV Measured From FDG PET is Associated With Outcome in Locally Advanced Cervical Cancer. Eur J Nucl Med Mol Imaging (2018) 45:187–95. doi: 10.1007/s00259-017-3824-z

PubMed Abstract | CrossRef Full Text | Google Scholar

49. Zhang L, Wu X, Liu J, Zhang B, Mo X, Chen Q, et al. MRI-Based Deep-Learning Model for Distant Metastasis-Free Survival in Locoregionally Advanced Nasopharyngeal Carcinoma. J Magn Reson Imaging (2021) 53:167–78. doi: 10.1002/jmri.27308

PubMed Abstract | CrossRef Full Text | Google Scholar

50. Fournier L, Costaridou L, Bidaut L, Michoux N, Lecouvet FE, de Geus-Oei LF, et al. Incorporating Radiomics Into Clinical Trials: Expert Consensus Endorsed by the European Society of Radiology on Considerations for Data-Driven Compared to Biologically Driven Quantitative Biomarkers. Eur Radiol (2021) 31:6001–12. doi: 10.1007/s00330-020-07598-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Keywords: cervical cancer, computed tomography, radiomics, nomogram, overall survival

Citation: Li H, Zhu M, Jian L, Bi F, Zhang X, Fang C, Wang Y, Wang J, Wu N and Yu X (2021) Radiomic Score as a Potential Imaging Biomarker for Predicting Survival in Patients With Cervical Cancer. Front. Oncol. 11:706043. doi: 10.3389/fonc.2021.706043

Received: 06 May 2021; Accepted: 19 July 2021;
Published: 16 August 2021.

Edited by:

Kathy Han, University Health Network, Canada

Reviewed by:

Federica Perelli, Santa Maria Annunziata Hospital, Italy
Fleur Huang, University of Alberta, Canada

Copyright © 2021 Li, Zhu, Jian, Bi, Zhang, Fang, Wang, Wang, Wu and Yu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Xiaoping Yu, yuxiaoping@hnca.org.cn; Nayiyuan Wu, wunayiyuan@163.com; Jing Wang, wangjing0081@hnszlyy.com

These authors have contributed equally to this work

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.