mpMRI-based habitat analysis for predicting prognoses in patients with high-grade serous ovarian cancer: a multicenter study

Qiu Bi, Kun Miao, Yang Liu, Jing Yang, Ao Zhou, Wenwei Shi, Lei Ying, Yunzhu Wu, Yang Song, Conghui Ai, Haiming Li, Jinwei Qiang

Abstract

Purpose: To evaluate the value of multiparametric MRI (mpMRI)-based habitat analysis for predicting prognoses in patients with high-grade serous ovarian cancer (HGSOC), and to develop combined models by integrating habitat analysis with clinical predictors.
Methods: This retrospective study included 503 HGSOC patients from four centers. A K-means algorithm was used to identify voxel clusters and generate habitats on mpMRI. Radiomics features were extracted from each habitat sub-region. After feature selection, habitat models were developed to predict overall survival (OS) and progression-free survival (PFS). Cox regression analyses were performed to identify clinical predictors and construct clinical models. Combined models were developed by integrating habitat signatures with clinical predictors.
Results: Compared with the clinical models (OS: 0.713 and 0.695; PFS: 0.727 and 0.700) and habitat models (OS: 0.707 and 0.672; PFS: 0.627 and 0.641), the combined models integrating habitat features and clinical independent predictors such as neoadjuvant chemotherapy (OS: 0.752 and 0.745; PFS: 0.784 and 0.754) achieved the highest C-indexes for predicting OS and PFS in the internal validation cohort and external test cohort.
Conclusion: The habitat models based on mpMRI demonstrated potential value in predicting the prognoses of HGSOC patients, but no significant advantages over the clinical models. The combined models were expected to improve the prognoses from the level of individual clinical characteristics and habitat features reflecting intratumoral heterogeneity.

Introduction

High-grade serous ovarian cancer (HGSOC) is the most lethal gynecological malignancy, a fact largely attributed to its extreme intratumoral heterogeneity. A single tumor is not a uniform mass; rather, it functions as a complex ecosystem containing various "neighborhoods" or habitats with distinct blood supplies, cell densities, and genetic profiles.

Standard clinical markers, such as FIGO stage or CA125 levels, often fail to capture this spatial complexity. This study addresses the need for more precise prognostic tools by utilizing habitat analysis—an unsupervised machine learning technique—to map these internal neighborhoods via multiparametric MRI (mpMRI). The goal was to determine if these specific spatial characteristics could better predict a patient's Overall Survival (OS) and Progression-Free Survival (PFS).

Methods

The study followed a retrospective, multicenter design involving 503 patients. The workflow was divided into four key stages:

  1. Habitat Generation: Researchers utilized four MRI sequences (T2WI, DWI, ADC, and CE-T1WI). They applied a K-means clustering algorithm to group voxels with similar signal intensities across these sequences, dividing each tumor into four distinct "habitats."
  2. Feature Extraction: From these four habitats, 27,888 radiomics features describing shape, texture, and intensity were extracted.
  3. Dimensionality Reduction: To prevent overfitting, the team used univariate Cox regression followed by LASSO-Cox selection to identify the most predictive features (45 for OS and 62 for PFS).
  4. Model Building: Three model types were constructed: a Clinical Model (based on factors like chemotherapy response), a Habitat Model (based purely on MRI sub-region features), and a Combined Model integrating both.

Datasets

The study utilized data from four different medical centers to ensure the model's robustness across different scanning environments and institutional protocols.

Table 1: Clinical characteristics of the study cohorts across training, internal validation, and external test sets.

The patient population was divided into three distinct cohorts:

  • Training Cohort (N=220): Used to develop the initial models.
  • Internal Validation Cohort (N=94): Used for initial verification of model accuracy.
  • External Test Cohort (N=189): Comprising patients from three separate centers to verify the model's generalizability.

Key Insights and Results

The study revealed that the internal "neighborhood" of the tumor matters significantly. The most predictive features were primarily derived from Habitat 2 (characterized by low ADC values and moderate enhancement), which likely represents the dense, solid, and most aggressive components of the tumor.

While habitat-only models were not significantly better than clinical models on their own, the Combined Model consistently outperformed all standalone versions across all metrics:

  • Overall Survival (OS): In the external test cohort, the Combined Model achieved a C-index of 0.745, compared to 0.695 for the clinical model and 0.672 for the habitat model.
  • Progression-Free Survival (PFS): The Combined Model reached a C-index of 0.754 in the external test, significantly higher than the clinical (0.700) or habitat (0.641) versions.
  • 3-Year Risk Prediction: The Area Under the Curve (AUC) for predicting death within three years was 0.821 for the combined model, demonstrating strong potential for identifying high-risk patients early.

Discussion and Limitations

The imaging data acted as a powerful "force multiplier," providing a granular view of tumor biology that clinical markers alone could not capture. The study confirmed that while neoadjuvant chemotherapy (NACT) and platinum resistance remain the strongest clinical predictors, MRI habitats provide the "missing link" to explain why patients with similar clinical profiles may have vastly different outcomes.

However, several limitations must be addressed before clinical implementation:

  • Manual Segmentation: Tumor segmentation was performed manually, which is a bottleneck for high-volume clinical settings. Automated, AI-driven segmentation is a necessary next step.
  • Retrospective Design: As a retrospective study, inherent selection bias exists. Prospective trials are required for real-time validation.
  • Biological Grounding: While habitats correlate with survival, the study did not perform "point-to-point" biopsies to confirm the exact cellular or genetic nature of each habitat (e.g., identifying regions of hypoxia vs. high proliferation).

In conclusion, mpMRI-based habitat analysis offers a promising avenue for improving HGSOC prognosis by accounting for the critical factor of intratumoral heterogeneity.

Summary updated

Abstract

<title>Abstract</title> Purpose To evaluate the value of multiparametric MRI (mpMRI)-based habitat analysis for predicting prognoses in patients with high-grade serous ovarian cancer (HGSOC), and to develop combined models by integrating habitat analysis with clinical predictors. Methods This retrospective study included 503 HGSOC patients from four centers. A K-means algorithm was used to identify voxel clusters and generate habitats on mpMRI. Radiomics features were extracted from each habitat sub-region. After feature selection, habitat models were developed to predict overall survival (OS) and progression-free survival (PFS). Cox regression analyses were performed to identify clinical predictors and construct clinical models. Combined models were developed by integrating habitat signatures with clinical predictors. Model performance was evaluated using C-index and time-dependent receiver operating characteristic area under the curves (AUCs). Results Compared with the clinical models (OS: 0.713 and 0.695; PFS: 0.727 and 0.700) and habitat models (OS: 0.707 and 0.672; PFS: 0.627 and 0.641), the combined models integrating habitat features and clinical independent predictors such as neoadjuvant chemotherapy (OS: 0.752 and 0.745; PFS: 0.784 and 0.754) achieved the highest C-indexes for predicting OS and PFS in the internal validation cohort and external test cohort. The combined models also achieved the highest AUCs in all cohorts. Conclusion The habitat models based on mpMRI demonstrated potential value in predicting the prognoses of HGSOC patients, but no significant advantages over the clinical models. The combined models were expected to improve the prognoses from the level of individual clinical characteristics and habitat features reflecting intratumoral heterogeneity.
BibTeX
@inproceedings{bi2025mpmri,
  author = {Qiu Bi and Kun Miao and Yang Liu and Jing Yang and Ao Zhou and Wenwei Shi and Lei Ying and Yunzhu Wu and Yang Song and Conghui Ai and Haiming Li and Jinwei Qiang},
  title = {{mpMRI-based habitat analysis for predicting prognoses in patients with high-grade serous ovarian cancer: a multicenter study}},
  booktitle = {Abdominal},
  year = {2025},
  doi = {10.1007/s00261-025-05004-9},
}
Figure