Radiomic profiles improve prognostication and reveal targets for therapy in cervical cancer.

Cervical cancer (CC) is a major global health problem with 570,000 new cases and 266,000 deaths annually. Prognosis is poor for advanced stage disease, and few effective treatments exist. Preoperative diagnostic imaging is common in high-income countries and MRI measured tumor size routinely guides treatment allocation of cervical cancer patients. Recently, the role of MRI radiomics has been recognized. However, its potential to independently predict survival and treatment response requires further clarification. This retrospective cohort study demonstrates how non-invasive, preoperative, MRI radiomic profiling may improve prognostication and tailoring of treatments and follow-ups for cervical cancer patients. By unsupervised clustering based on 293 radiomic features from 132 patients, we identify three distinct clusters comprising patients with significantly different risk profiles, also when adjusting for FIGO stage and age. By linking their radiomic profiles to genomic alterations, we identify putative treatment targets for the different patient clusters (e.g., immunotherapy, CDK4/6 and YAP-TEAD inhibitors and p53 pathway targeting treatments).

Clinicopathological and radiological stratification within FIGO 2018 stages improves risk-prediction in cervical cancer.

OBJECTIVE: Assess the added prognostic value of the updated International Federation of Gynecology and Obstetrics (FIGO) 2018 staging system, and to identify clinicopathological and radiological biomarkers for improved FIGO 2018 prognostication. METHODS: Patient data were retrieved from a prospectively collected patient cohort including all consenting patients with cervical cancer diagnosed and treated at Haukeland University Hospital during 2001-2022 (n = 948). All patients were staged according to the FIGO 2009 and FIGO 2018 guidelines based on available data for individual patients. MRI-assessed maximum tumor diameter and stromal tumor invasion, as well as histopathologically assessed lymphovascular space invasion were applied to categorize patients according to the Sedlis criteria. RESULTS: FIGO 2018 stage yielded the highest area under the receiver operating characteristic (ROC) curve (AUC) (0.86 versus 0.81 for FIGO 2009) for predicting disease-specific survival. The most common stage migration in FIGO 2018 versus FIGO 2009 was upstaging from stages IB/II to stage IIIC due to suspicious lymph nodes identified by PET/CT and/or MRI. In FIGO 2018 stage III patients, extent and size of primary tumor (p = 0.04), as well as its histological type (p = 0.003) were highly prognostic. Sedlis criteria were prognostic within FIGO 2018 IB patients (p = 0.04). CONCLUSIONS: Incorporation of cross-sectional imaging increases prognostic precision, as suggested by the FIGO 2018 guidelines. The 2018 FIGO IIIC stage could be refined by including the size and extent of primary tumor and histological type. The FIGO IB risk prediction could be improved by applying MRI-assessed tumor size and stromal invasion.

MRI-based radiomic signatures for pretreatment prognostication in cervical cancer.

BACKGROUND: Accurate pretherapeutic prognostication is important for tailoring treatment in cervical cancer (CC). PURPOSE: To investigate whether pretreatment MRI-based radiomic signatures predict disease-specific survival (DSS) in CC. STUDY TYPE: Retrospective. POPULATION: CC patients (n = 133) allocated into training(T) (nT = 89)/validation(V) (nV = 44) cohorts. FIELD STRENGTH/SEQUENCE: T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI) at 1.5T or 3.0T. ASSESSMENT: Radiomic features from segmented tumors were extracted from T2WI and DWI (high b-value DWI and apparent diffusion coefficient (ADC) maps). STATISTICAL TESTS: Radiomic signatures for prediction of DSS from T2WI (T2rad ) and T2WI with DWI (T2 + DWIrad ) were constructed by least absolute shrinkage and selection operator (LASSO) Cox regression. Area under time-dependent receiver operating characteristics curves (AUC) were used to evaluate and compare the prognostic performance of the radiomic signatures, MRI-derived maximum tumor size ≤/> 4 cm (MAXsize ), and 2018 International Federation of Gynecology and Obstetrics (FIGO) stage (I-II/III-IV). Survival was analyzed using Cox model estimating hazard ratios (HR) and Kaplan-Meier method with log-rank tests. RESULTS: The radiomic signatures T2rad and T2 + DWIrad yielded AUCT /AUCV of 0.80/0.62 and 0.81/0.75, respectively, for predicting 5-year DSS. Both signatures yielded better or equal prognostic performance to that of MAXsize (AUCT /AUCV : 0.69/0.65) and FIGO (AUCT /AUCV : 0.77/0.64) and were significant predictors of DSS after adjusting for FIGO (HRT /HRV for T2rad : 4.0/2.5 and T2 + DWIrad : 4.8/2.1). Adding T2rad and T2 + DWIrad to FIGO significantly improved DSS prediction compared to FIGO alone in cohort(T) (AUCT 0.86 and 0.88 vs. 0.77), and FIGO with T2 + DWIrad tended to the same in cohort(V) (AUCV 0.75 vs. 0.64, p = 0.07). High radiomic score for T2 + DWIrad was significantly associated with reduced DSS in both cohorts. DATA CONCLUSION: Radiomic signatures from T2WI and T2WI with DWI may provide added value for pretreatment risk assessment and for guiding tailored treatment strategies in CC.

Preoperative pelvic MRI and 2-[18F]FDG PET/CT for lymph node staging and prognostication in endometrial cancer-time to revisit current imaging guidelines?

OBJECTIVE: This study presents the diagnostic performance of four different preoperative imaging workups (IWs) for prediction of lymph node metastases (LNMs) in endometrial cancer (EC): pelvic MRI alone (IW1), MRI and [18F]FDG-PET/CT in all patients (IW2), MRI with selective [18F]FDG-PET/CT if high-risk preoperative histology (IW3), and MRI with selective [18F]FDG-PET/CT if MRI indicates FIGO stage ≥ 1B (IW4). METHODS: In 361 EC patients, preoperative staging parameters from both pelvic MRI and [18F]FDG-PET/CT were recorded. Area under receiver operating characteristic curves (ROC AUC) compared the diagnostic performance for the different imaging parameters and workups for predicting surgicopathological FIGO stage. Survival data were assessed using Kaplan-Meier estimator with log-rank test. RESULTS: MRI and [18F]FDG-PET/CT staging parameters yielded similar AUCs for predicting corresponding FIGO staging parameters in low-risk versus high-risk histology groups (p ≥ 0.16). The sensitivities, specificities, and AUCs for LNM prediction were as follows: IW1-33% [9/27], 95% [185/193], and 0.64; IW2-56% [15/27], 90% [174/193], and 0.73 (p = 0.04 vs. IW1); IW3-44% [12/27], 94% [181/193], and 0.69 (p = 0.13 vs. IW1); and IW4-52% [14/27], 91% [176/193], and 0.72 (p = 0.06 vs. IW1). IW3 and IW4 selected 34% [121/361] and 54% [194/361] to [18F]FDG-PET/CT, respectively. Employing IW4 identified three distinct patient risk groups that exhibited increasing FIGO stage (p < 0.001) and stepwise reductions in survival (p ≤ 0.002). CONCLUSION: Selective [18F]FDG-PET/CT in patients with high-risk MRI findings yields better detection of LNM than MRI alone, and similar diagnostic performance to that of MRI and [18F]FDG-PET/CT in all. KEY POINTS: • Imaging by MRI and [18F]FDG PET/CT yields similar diagnostic performance in low- and high-risk histology groups for predicting central FIGO staging parameters. • Utilizing a stepwise imaging workup with MRI in all patients and [18F]FDG-PET/CT in selected patients based on MRI findings identifies preoperative risk groups exhibiting significantly different survival. • The proposed imaging workup selecting ~54% of the patients to [18F]FDG-PET/CT yield better detection of LNMs than MRI alone, and similar LNM detection to that of MRI and [18F]FDG-PET/CT in all.

What MRI-based tumor size measurement is best for predicting long-term survival in uterine cervical cancer?

BACKGROUND: Tumor size assessment by MRI is central for staging uterine cervical cancer. However, the optimal role of MRI-derived tumor measurements for prognostication is still unclear. MATERIAL AND METHODS: This retrospective cohort study included 416 women (median age: 43 years) diagnosed with cervical cancer during 2002-2017 who underwent pretreatment pelvic MRI. The MRIs were independently read by three radiologists, measuring maximum tumor diameters in three orthogonal planes and maximum diameter irrespective of plane (MAXimaging). Inter-reader agreement for tumor size measurements was assessed by intraclass correlation coefficients (ICCs). Size was analyzed in relation to age, International Federation of Gynecology and Obstetrics (FIGO) (2018) stage, histopathological markers, and disease-specific survival using Kaplan-Meier-, Cox regression-, and time-dependent receiver operating characteristics (tdROC) analyses. RESULTS: All MRI tumor size variables (cm) yielded high areas under the tdROC curves (AUCs) for predicting survival (AUC 0.81-0.84) at 5 years after diagnosis and predicted outcome (hazard ratios [HRs] of 1.42-1.76, p < 0.001 for all). Only MAXimaging independently predicted survival (HR = 1.51, p = 0.03) in the model including all size variables. The optimal cutoff for maximum tumor diameter (≥ 4.0 cm) yielded sensitivity (specificity) of 83% (73%) for predicting disease-specific death after 5 years. Inter-reader agreement for MRI-based primary tumor size measurements was excellent, with ICCs of 0.83-0.85. CONCLUSION: Among all MRI-derived tumor size measurements, MAXimaging was the only independent predictor of survival. MAXimaging ≥ 4.0 cm represents the optimal cutoff for predicting long-term disease-specific survival in cervical cancer. Inter-reader agreement for MRI-based tumor size measurements was excellent.

Interobserver agreement and prognostic impact for MRI-based 2018 FIGO staging parameters in uterine cervical cancer.

OBJECTIVES: To evaluate the interobserver agreement for MRI-based 2018 International Federation of Gynecology and Obstetrics (FIGO) staging parameters in patients with cervical cancer and assess the prognostic value of these MRI parameters in relation to other clinicopathological markers. METHODS: This retrospective study included 416 women with histologically confirmed cervical cancer who underwent pretreatment pelvic MRI from May 2002 to December 2017. Three radiologists independently recorded MRI-derived staging parameters incorporated in the 2018 FIGO staging system. Kappa coefficients (κ) for interobserver agreement were calculated. The predictive and prognostic values of the MRI parameters were explored using ROC analyses and Kaplan-Meier with log-rank tests, and analyzed in relation to clinicopathological patient characteristics. RESULTS: Overall agreement was substantial for the staging parameters: tumor size > 2 cm (κ = 0.80), tumor size > 4 cm (κ = 0.76), tumor size categories (≤ 2 cm; > 2 and ≤ 4 cm; > 4 cm) (κ = 0.78), parametrial invasion (κ = 0.63), vaginal invasion (κ = 0.61), and enlarged lymph nodes (κ = 0.63). Higher MRI-derived tumor size category (≤ 2 cm; > 2 and ≤ 4 cm; > 4 cm) was associated with a stepwise reduction in survival (p ≤ 0.001 for all). Tumor size > 4 cm and parametrial invasion at MRI were associated with aggressive clinicopathological features, and the incorporation of these MRI-based staging parameters improved risk stratification when compared to corresponding clinical assessments alone. CONCLUSION: The interobserver agreement for central MRI-derived 2018 FIGO staging parameters was substantial. MRI improved the identification of patients with aggressive clinicopathological features and poor survival, demonstrating the potential impact of MRI enabling better prognostication and treatment tailoring in cervical cancer. KEY POINTS: • The overall interobserver agreement was substantial (κ values 0.61-0.80) for central MRI staging parameters in the 2018 FIGO system. • Higher MRI-derived tumor size category was linked to a stepwise reduction in survival (p ≤ 0.001 for all). • MRI-derived tumor size > 4 cm and parametrial invasion were associated with aggressive clinicopathological features, and the incorporation of these MRI-derived staging parameters improved risk stratification when compared to clinical assessments alone.

A radiogenomics application for prognostic profiling of endometrial cancer.

Prognostication is critical for accurate diagnosis and tailored treatment in endometrial cancer (EC). We employed radiogenomics to integrate preoperative magnetic resonance imaging (MRI, n = 487 patients) with histologic-, transcriptomic- and molecular biomarkers (n = 550 patients) aiming to identify aggressive tumor features in a study including 866 EC patients. Whole-volume tumor radiomic profiling from manually (radiologists) segmented tumors (n = 138 patients) yielded clusters identifying patients with high-risk histological features and poor survival. Radiomic profiling by a fully automated machine learning (ML)-based tumor segmentation algorithm (n = 336 patients) reproduced the same radiomic prognostic groups. From these radiomic risk-groups, an 11-gene high-risk signature was defined, and its prognostic role was reproduced in orthologous validation cohorts (n = 554 patients) and aligned with The Cancer Genome Atlas (TCGA) molecular class with poor survival (copy-number-high/p53-altered). We conclude that MRI-based integrated radiogenomics profiling provides refined tumor characterization that may aid in prognostication and guide future treatment strategies in EC.

A 10-gene prognostic signature points to LIMCH1 and HLA-DQB1 as important players in aggressive cervical cancer disease.

BACKGROUND: Advanced cervical cancer carries a particularly poor prognosis, and few treatment options exist. Identification of effective molecular markers is vital to improve the individualisation of treatment. We investigated transcriptional data from cervical carcinomas related to patient survival and recurrence to identify potential molecular drivers for aggressive disease. METHODS: Primary tumour RNA-sequencing profiles from 20 patients with recurrence and 53 patients with cured disease were compared. Protein levels and prognostic impact for selected markers were identified by immunohistochemistry in a population-based patient cohort. RESULTS: Comparison of tumours relative to recurrence status revealed 121 differentially expressed genes. From this gene set, a 10-gene signature with high prognostic significance (p = 0.001) was identified and validated in an independent patient cohort (p = 0.004). Protein levels of two signature genes, HLA-DQB1 (n = 389) and LIMCH1 (LIM and calponin homology domain 1) (n = 410), were independent predictors of survival (hazard ratio 2.50, p = 0.007 for HLA-DQB1 and 3.19, p = 0.007 for LIMCH1) when adjusting for established prognostic markers. HLA-DQB1 protein expression associated with programmed death ligand 1 positivity (p < 0.001). In gene set enrichment analyses, HLA-DQB1high tumours associated with immune activation and response to interferon-γ (IFN-γ). CONCLUSIONS: This study revealed a 10-gene signature with high prognostic power in cervical cancer. HLA-DQB1 and LIMCH1 are potential biomarkers guiding cervical cancer treatment.

Automated segmentation of endometrial cancer on MR images using deep learning.

Preoperative MR imaging in endometrial cancer patients provides valuable information on local tumor extent, which routinely guides choice of surgical procedure and adjuvant therapy. Furthermore, whole-volume tumor analyses of MR images may provide radiomic tumor signatures potentially relevant for better individualization and optimization of treatment. We apply a convolutional neural network for automatic tumor segmentation in endometrial cancer patients, enabling automated extraction of tumor texture parameters and tumor volume. The network was trained, validated and tested on a cohort of 139 endometrial cancer patients based on preoperative pelvic imaging. The algorithm was able to retrieve tumor volumes comparable to human expert level (likelihood-ratio test, [Formula: see text]). The network was also able to provide a set of segmentation masks with human agreement not different from inter-rater agreement of human experts (Wilcoxon signed rank test, [Formula: see text], [Formula: see text], and [Formula: see text]). An automatic tool for tumor segmentation in endometrial cancer patients enables automated extraction of tumor volume and whole-volume tumor texture features. This approach represents a promising method for automatic radiomic tumor profiling with potential relevance for better prognostication and individualization of therapeutic strategy in endometrial cancer.