Anatomy-informed Data Augmentation for Enhanced Prostate Cancer Detection

Balint Kovacs,Nils Netzer,Michael Baumgartner,Carolin Eith,Dimitrios Bounias,Clara Meinzer,Paul F. Jaeger,Kevin S. Zhang,Ralf Floca,Adrian Schrader,Fabian Isensee,Regula Gnirs,Magdalena Goertz,Viktoria Schuetz,Albrecht Stenzinger,Markus Hohenfellner,Heinz-Peter Schlemmer,Ivo Wolf,David Bonekamp,Klaus H. Maier-Hein

2023-09-07

Abstract:Data augmentation (DA) is a key factor in medical image analysis, such as in prostate cancer (PCa) detection on magnetic resonance images. State-of-the-art computer-aided diagnosis systems still rely on simplistic spatial transformations to preserve the pathological label post transformation. However, such augmentations do not substantially increase the organ as well as tumor shape variability in the training set, limiting the model's ability to generalize to unseen cases with more diverse localized soft-tissue deformations. We propose a new anatomy-informed transformation that leverages information from adjacent organs to simulate typical physiological deformations of the prostate and generates unique lesion shapes without altering their label. Due to its lightweight computational requirements, it can be easily integrated into common DA frameworks. We demonstrate the effectiveness of our augmentation on a dataset of 774 biopsy-confirmed examinations, by evaluating a state-of-the-art method for PCa detection with different augmentation settings.

Image and Video Processing,Artificial Intelligence,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper primarily focuses on data augmentation techniques in prostate cancer detection, aiming to improve the detection performance in magnetic resonance imaging (MRI) through a novel anatomy-informed data augmentation method. ### Research Background and Objectives Currently, in medical image analysis, especially for prostate cancer detection, Data Augmentation (DA) is a key technology. Existing computer-aided diagnosis systems typically rely on simple spatial transformations to maintain the accuracy of lesion labels after transformation. However, these methods do not substantially increase the variability of organ and tumor shapes in the training set, limiting the model's ability to generalize to more diverse local soft tissue deformations in unseen cases. ### Proposed Method The paper proposes a new anatomy-informed transformation method that uses information from adjacent organs to simulate typical physiological deformations and generate unique lesion shapes without altering their labels. This method is computationally lightweight and can be easily integrated into commonly used data augmentation frameworks. ### Experimental Results - **Visual Assessment**: Rigorous Turing tests indicate that the proposed anatomy-informed deformation produces highly realistic soft tissue deformations. - **Quantitative Assessment**: - In patient-level prostate cancer diagnosis, the use of anatomy-informed data augmentation improved sensitivity, approaching the level of radiologists. - In lesion-level prostate cancer detection, this method also performed excellently, significantly enhancing performance compared to basic data augmentation schemes. ### Conclusion In summary, the paper proposes a novel and practical anatomy-informed data augmentation method that can simulate typical soft tissue deformations, effectively improving prostate cancer detection performance, making it closer to the professional level of clinicians. Additionally, due to its computational simplicity, this method is easy to integrate into existing data augmentation frameworks, showing high application potential.

Anatomy-informed Data Augmentation for Enhanced Prostate Cancer Detection

A Comprehensive Study of Data Augmentation Strategies for Prostate Cancer Detection in Diffusion-Weighted MRI Using Convolutional Neural Networks

Addressing image misalignments in multi-parametric prostate MRI for enhanced computer-aided diagnosis of prostate cancer

Artifact Augmentation for Enhanced Tissue Detection in Microscope Scanner Systems

Synthetic Histology Images for Training AI Models: A Novel Approach to Improve Prostate Cancer Diagnosis

Data Augmentation and Transfer Learning to Improve Generalizability of an Automated Prostate Segmentation Model.

Dynamic Data Augmentation via MCTS for Prostate MRI Segmentation

A statistical deformation model-based data augmentation method for volumetric medical image segmentation

Morphological Change Forecasting For Prostate Glands Using Feature-Based Registration And Kernel Density Extrapolation

Mask-guided Data Augmentation for Multiparametric MRI Generation with a Rare Hepatocellular Carcinoma

Automatic detection of prostate cancer grades and chronic prostatitis in biparametric MRI

Latent Diffusion Models with Image-Derived Annotations for Enhanced AI-Assisted Cancer Diagnosis in Histopathology

Data Augmentation for Surgical Scene Segmentation with Anatomy-Aware Diffusion Models

Annotation-efficient cancer detection with report-guided lesion annotation for deep learning-based prostate cancer detection in bpMRI

Overcoming Limitations in Artificial Intelligence-based Prostate Cancer Detection through Better Datasets and a Bayesian Approach to Aggregate Panel Predictions

Protecting Prostate Cancer Classification From Rectal Artifacts via Targeted Adversarial Training

Evaluation of automatic discrimination between benign and malignant prostate tissue in the era of high precision digital pathology

The Effectiveness of Data Augmentation for Detection of Gastrointestinal Diseases from Endoscopical Images

Exploring synthetic datasets for computer-aided detection: a case study using phantom scan data for enhanced lung nodule false positive reduction

AutoPET Challenge: Tumour Synthesis for Data Augmentation

Deformable MRI Sequence Registration for AI-based Prostate Cancer Diagnosis