Generating Counterfactual Explanations For Causal Inference in Breast Cancer Treatment Response

Zhou, Siqiong; Pfeiffer, Nicholaus; Islam, Upala J.; Banerjee, Imon; Patel, Bhavika; Iquebal, Ashif Sikandar

doi:10.1109/case49997.2022.9926519

Cited by 3 publications

(8 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After comparing our results with those in [23], we found that our proposed SCGAN method is unable to generate counterfactual instances with only one feature change, using the model built over all 536 features. However, it is worth noting that the 10-feature model used in [23] was much simpler than our current model and therefore may not have captured all the relevant factors that contribute to the predicted outcome. In contrast, our SCGAN model is more complex and takes into account interactions between multiple features, making the resulting counterfactuals more realistic.…”

Section: Case Study: Breast Cancer Dce-mrimentioning

confidence: 88%

“…Counterfactual explanations can provide a deeper understanding of the causal relationships between various medical factors, helping clinicians to make more informed decisions. In our recent work, Zhou et al [23] propose a DiCE-based method for identifying causal relationships between imaging phenotypes, clinical information, molecular features, and treatment response in breast cancer patients. The authors compare their approach to traditional explanation methods, such as LIME and Shapley, and highlight the advantages of the counterfactual approach.…”

Section: Background and Related Workmentioning

confidence: 99%

“…We finally apply our method to the open-source pre-operative DCE-MRI data from Saha et al, which extracted 529 radiomic features for 922 breast cancer patients diagnosed with invasive breast cancer between January 2000 to March 2014 [46]. Please refer to [27] for details on the radiomic features. Out of the 922 patients, 288 were evaluated for neoadjuvant therapy outcomes, categorized as either pCR (64 patients) or pNR (224 patients).…”

Section: Numerical Experimentsmentioning

confidence: 99%

“…Out of the 922 patients, 288 were evaluated for neoadjuvant therapy outcomes, categorized as either pCR (64 patients) or pNR (224 patients). To handle imbalance in the breast cancer dataset, random undersampling is applied as discussed in [27].…”

Section: Numerical Experimentsmentioning

confidence: 99%

“…Our approach employed DiCE to generate counterfactuals. However, due to the computational cost of DiCE, we preselected the top 10 features using Shapley values to generate the counterfactual instances [23]. Limiting the feature space significantly limited the ability of our method to investigate the feature space that comprised 536 features (see Section 4.3).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

SCGAN: Sparse CounterGAN for Counterfactual Explanations in Breast Cancer Prediction

Zhou

Islam

Pfeiffer

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Imaging phenotypes extracted via radiomics of magnetic resonance imaging have shown great potential in predicting the treatment response in breast cancer patients after administering neoadjuvant systemic therapy (NST). Understanding the causal relationships between Imaging phenotypes, Clinical information, and Molecular (ICM) features, and the treatment response are critical in guiding treatment strategies and management plans. Counterfactual explanations provide an interpretable approach to generating causal inference; however, existing approaches are either computationally prohibitive for high dimensional problems, generate unrealistic counterfactuals, or confound the effects of causal features. This paper proposes a new method called Sparse CounteRGAN (SCGAN) for generating counterfactual instances to establish causal relationships between ICM features and the treatment response after NST. The generative approach learns the distribution of the original instances and, therefore, ensures that the new instances are realistic. Further, we propose a loss function that regularizes the counterfactuals to minimize the distance between original instances and counterfactuals (to promote sparsity) and the distances among generated counterfactuals to promote diversity. We evaluate the proposed method on two publicly available datasets, followed by the breast cancer dataset, and compare their performance with existing methods in the literature. Finally, we demonstrate the causal relationships from generated counterfactual instances. Results show that SCGAN generates plausible and realistic counterfactual instances with small changes in only a few features, making it a valuable tool for understanding the causal relationships between ICM features and treatment response.

show abstract

Section: Case Study: Breast Cancer Dce-mrimentioning

confidence: 88%

Section: Background and Related Workmentioning

confidence: 99%

Section: Numerical Experimentsmentioning

confidence: 99%

Section: Numerical Experimentsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

SCGAN: Sparse CounterGAN for Counterfactual Explanations in Breast Cancer Prediction

Zhou

Islam

Pfeiffer

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

A Framework for Interpretability in Machine Learning for Medical Imaging

Wang,

Karaman,

Kim

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Interpretability for machine learning models in medical imaging (MLMI) is an important direction of research. However, there is a general sense of murkiness in what interpretability means. Why does the need for interpretability in MLMI arise? What goals does one actually seek to address when interpretability is needed? To answer these questions, we identify a need to formalize the goals and elements of interpretability in MLMI. By reasoning about real-world tasks and goals common in both medical image analysis and its intersection with machine learning, we identify five core elements of interpretability: localization, visual recognizability, physical attribution, model transparency, and actionability. From this, we arrive at a framework for interpretability in MLMI, which serves as a step-by-step guide to approaching interpretability in this context. Overall, this paper formalizes interpretability needs in the context of medical imaging, and our applied perspective clarifies concrete MLMI-specific goals and considerations in order to guide method design and improve real-world usage. Our goal is to provide practical and didactic information for model designers and practitioners, inspire developers of models in the medical imaging field to reason more deeply about what interpretability is achieving, and suggest future directions of interpretability research.

show abstract

SCGAN: Sparse CounterGAN for Counterfactual Explanations in Breast Cancer Prediction

Zhou,

Islam,

Pfeiffer

et al. 2024

IEEE Trans. Automat. Sci. Eng.

View full text Add to dashboard Cite

Imaging phenotypes extracted via radiomics of magnetic resonance imaging have shown great potential in predicting the treatment response in breast cancer patients after administering neoadjuvant systemic therapy (NST). Understanding the causal relationships between the treatment response and Imaging phenotypes, Clinical information, and Molecular (ICM) features are critical in guiding treatment strategies and management plans. Counterfactual explanations provide an interpretable approach to generating causal inference. However, existing approaches are either computationally prohibitive for high dimensional problems, generate unrealistic counterfactuals, or confound the effects of causal features by changing multiple features simultaneously. This paper proposes a new method called Sparse Coun-teRGAN (SCGAN) for generating counterfactual instances to reveal causal relationships between ICM features and the treatment response after NST. The generative approach learns the distribution of the original instances and, therefore, ensures that the new instances are realistic. We propose dropout training of the discriminator to promote sparsity and introduce a diversity term in the loss function to maximize the distances among generated counterfactuals. We evaluate the proposed method on two publicly available datasets, followed by the breast cancer dataset, and compare their performance with existing methods in the literature. Results show that SCGAN generates sparse and diverse counterfactual instances that also achieve plausibility and feasibility, making it a valuable tool for understanding the causal relationships between ICM features and treatment response.

show abstract

Generating Counterfactual Explanations For Causal Inference in Breast Cancer Treatment Response

Cited by 3 publications

References 0 publications

SCGAN: Sparse CounterGAN for Counterfactual Explanations in Breast Cancer Prediction

SCGAN: Sparse CounterGAN for Counterfactual Explanations in Breast Cancer Prediction

A Framework for Interpretability in Machine Learning for Medical Imaging

SCGAN: Sparse CounterGAN for Counterfactual Explanations in Breast Cancer Prediction

Contact Info

Product

Resources

About