Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

APIS: A paired CT-MRI dataset for ischemic stroke segmentation challenge

Santiago Gómez, Daniel Mantilla, Gustavo Garzón, Edgar Rangel, Andrés Ortiz, Franklin Sierra-Jerez, Fabio Martínez

TL;DR

The paper introduces APIS, the first public paired NCCT–ADC dataset for ischemic stroke segmentation and a related ISBI 2023 challenge to evaluate CT-based lesion delineation with diffusion MRI guidance. It demonstrates radiologist agreement on ADC lesions and shows markedly better segmentation performance when ADC is available, compared with NCCT alone. The results underscore the value of cross-modality information and motivate future work on expanding paired data, robust registration, and synthetic modality generation to improve generalization. Overall, APIS provides a foundation for multimodal stroke segmentation research and benchmarking of image-to-image translation and integration approaches.

Abstract

Stroke is the second leading cause of mortality worldwide. Immediate attention and diagnosis play a crucial role regarding patient prognosis. The key to diagnosis consists in localizing and delineating brain lesions. Standard stroke examination protocols include the initial evaluation from a non-contrast CT scan to discriminate between hemorrhage and ischemia. However, non-contrast CTs may lack sensitivity in detecting subtle ischemic changes in the acute phase. As a result, complementary diffusion-weighted MRI studies are captured to provide valuable insights, allowing to recover and quantify stroke lesions. This work introduced APIS, the first paired public dataset with NCCT and ADC studies of acute ischemic stroke patients. APIS was presented as a challenge at the 20th IEEE International Symposium on Biomedical Imaging 2023, where researchers were invited to propose new computational strategies that leverage paired data and deal with lesion segmentation over CT sequences. Despite all the teams employing specialized deep learning tools, the results suggest that the ischemic stroke segmentation task from NCCT remains challenging. The annotated dataset remains accessible to the public upon registration, inviting the scientific community to deal with stroke characterization from NCCT but guided with paired DWI information.

APIS: A paired CT-MRI dataset for ischemic stroke segmentation challenge

TL;DR

The paper introduces APIS, the first public paired NCCT–ADC dataset for ischemic stroke segmentation and a related ISBI 2023 challenge to evaluate CT-based lesion delineation with diffusion MRI guidance. It demonstrates radiologist agreement on ADC lesions and shows markedly better segmentation performance when ADC is available, compared with NCCT alone. The results underscore the value of cross-modality information and motivate future work on expanding paired data, robust registration, and synthetic modality generation to improve generalization. Overall, APIS provides a foundation for multimodal stroke segmentation research and benchmarking of image-to-image translation and integration approaches.

Abstract

Stroke is the second leading cause of mortality worldwide. Immediate attention and diagnosis play a crucial role regarding patient prognosis. The key to diagnosis consists in localizing and delineating brain lesions. Standard stroke examination protocols include the initial evaluation from a non-contrast CT scan to discriminate between hemorrhage and ischemia. However, non-contrast CTs may lack sensitivity in detecting subtle ischemic changes in the acute phase. As a result, complementary diffusion-weighted MRI studies are captured to provide valuable insights, allowing to recover and quantify stroke lesions. This work introduced APIS, the first paired public dataset with NCCT and ADC studies of acute ischemic stroke patients. APIS was presented as a challenge at the 20th IEEE International Symposium on Biomedical Imaging 2023, where researchers were invited to propose new computational strategies that leverage paired data and deal with lesion segmentation over CT sequences. Despite all the teams employing specialized deep learning tools, the results suggest that the ischemic stroke segmentation task from NCCT remains challenging. The annotated dataset remains accessible to the public upon registration, inviting the scientific community to deal with stroke characterization from NCCT but guided with paired DWI information.
Paper Structure (6 sections, 5 equations, 3 figures, 4 tables)

This paper contains 6 sections, 5 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Data
  3. Challenge
  4. Results
  5. Discussion
  6. Conclusions

Figures (3)

  • Figure 1: Image modalities in the APIS dataset, from left to right: NCCT and MRI-ADC. Second column shows a high (top) and low (bottom) level of agreement over segmentation.
  • Figure 2: Models performance against expert annotations from NCCT sources. From left to right, the R1 lesion identification, R1 y R2 identification agreement, R2 lesion identification, lesion identification over ADC but not over CT source, and finally, patients control annotations for both experts and R2, respectively.
  • Figure 3: Models performance against expert annotations from ADC sources. From left to right, the R1 lesion identification, R1 y R2 identification agreement, R2 lesion identification, lesion identification over ADC but not over CT source, and finally, patients control annotations for both experts and R2, respectively.