Journals / Firat University Journal of Experimental and Computational Engineering / 2026 / Cilt: 5 Sayı: 2

Beyond Accuracy: A Clinically-Oriented Breast Cancer Diagnostic Framework Integrating Conformal Prediction, Counterfactual Explanations, and Multi-Task Learning on FNA Cytological Features

Doğruluğun Ötesinde: FNA Sitolojik Özellikleri Üzerinde Konformal Tahmin, Karşıt-Olgusal Açıklamalar ve Çok Görevli Öğrenmeyi Entegre Eden Klinik Odaklı Bir Meme Kanseri Tanı Çerçevesi

Esra Yüce(YOZGAT BOZOK UNIVERSITY)0000-0002-9522-8352
Enes Eren Süzgen(YOZGAT BOZOK UNIVERSITY)0009-0001-5442-930X
Muhammet Emin Şahin(IZMIR BAKIRCAY UNIVERSITY)0000-0001-7729-990X
Pages
552–573
Publication Date
2026-06-27
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

Finally, multi-task analysis has shown that cytological morphology is useful for the purposes of breast cancer diagnosis but not prognosis.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:Breast cancer remains a serious global health threat, and there is a clear need for diagnostically reliable research frameworks that go beyond the overly optimistic performance claims often reported in the literature.
📊
Key Findings & Data:The methodology presented uses Wisconsin Diagnostic Breast Cancer and Wisconsin Prognostic Breast Cancer datasets to develop a completely leak-proof nested 5x5 cross-validation procedure where any preprocessing was limited to training folds only.
💡
Conclusion & Impact:Finally, multi-task analysis has shown that cytological morphology is useful for the purposes of breast cancer diagnosis but not prognosis.
🔬Design:Araştırma Makalesi
📍Scope:Comprehensive Sample
⚙️Variables:Meme kanseri · Konformal tahmin. Karşıt-olgusal açıklamalar · İstifleme topluluğu

Abstract

Breast cancer remains a serious global health threat, and there is a clear need for diagnostically reliable research frameworks that go beyond the overly optimistic performance claims often reported in the literature. This study provides a clinical diagnostic methodology for breast cancer using fine-needle aspiration cytology, which is intended to promote clinical validation beyond benchmark maximization. The methodology presented uses Wisconsin Diagnostic Breast Cancer and Wisconsin Prognostic Breast Cancer datasets to develop a completely leak-proof nested 5x5 cross-validation procedure where any preprocessing was limited to training folds only. The original 30 cytology features were increased to 42 by adding biologically-motivated feature engineering, and a stacked ensemble method was trained to provide clinically interpretable probability estimates. The method used bootstrap-based feature stability, split conformal prediction to perform uncertainty-driven triage, feature attribution and counterfactual explanation modules, decision curve analysis for clinical validation, and multi-task learning for cross-dataset transfer analysis. In addition, the method yielded an out-of-fold area under the curve of 0.9932 and a holdout accuracy of 0.9912. Given 95% confidence interval, split conformal prediction resulted in a definitive single-class prediction for 99.1% of instances, suggesting that uncertain cases constitute only a negligible number and require only an escalation in clinical practice. The results of decision curve analysis have shown that there is always net clinical benefit across all clinically relevant probability thresholds. Finally, multi-task analysis has shown that cytological morphology is useful for the purposes of breast cancer diagnosis but not prognosis.

Özet

Meme kanseri küresel ölçekte ciddi bir sağlık tehdidi olmaya devam etmektedir ve literatürde sıklıkla bildirilen aşırı iyimser performans iddialarının ötesine geçen, tanısal açıdan güvenilir araştırma çerçevelerine açık bir ihtiyaç vardır. Bu çalışma, kıyaslama maksimizasyonunun ötesinde klinik doğrulamayı teşvik etmek amacıyla ince iğne aspirasyon sitolojisini kullanan bir meme kanseri klinik tanı metodolojisi ortaya koymaktadır. Sunulan metodoloji, yalnızca eğitim katlamalarıyla sınırlı tutulan ön işleme adımlarıyla tamamen sızıntısız, iç içe geçmiş 5x5 çapraz doğrulama prosedürü geliştirmek için Wisconsin Tanısal Meme Kanseri ve Wisconsin Prognostik Meme Kanseri veri setlerini kullanmaktadır. Orijinal 30 sitoloji özelliği, biyolojik temelli özellik mühendisliği yoluyla 42'ye çıkarılmış ve klinik olarak yorumlanabilir olasılık tahminleri sunmak amacıyla yığılmış bir topluluk yöntemi eğitilmiştir. Yöntemde; önyükleme tabanlı özellik kararlılığı, belirsizlik güdümlü triyaj gerçekleştirmek için bölünmüş konformal tahmin, özellik atıf ve karşıt olgusal açıklama modülleri, klinik doğrulama için karar eğrisi analizi ve çapraz veri seti aktarım analizi için çok görevli öğrenme kullanılmıştır. Bunlara ek olarak, yöntem 0.9932 katlama dışı eğri altı alan ve 0.9912 tutma doğruluğu değerleri elde etmiştir. %95 güven aralığında bölünmüş konformal tahmin, örneklerin %99,1'i için kesin tek sınıf tahmini üretmiş; bu da belirsiz vakaların yalnızca ihmal edilebilir bir oranı oluşturduğunu ve klinik pratikte yalnızca bir yükseltme gerektirdiğini ortaya koymaktadır. Karar eğrisi analizinin sonuçları, klinik açıdan ilgili tüm olasılık eşiklerinde sürekli olarak net klinik yarar bulunduğunu göstermiştir. Son olarak, çok görevli analiz, sitolojik morfolojinin meme kanseri tanısı açısından yararlı olduğunu, ancak prognoz açısından bu niteliği taşımadığını ortaya koymuştur.

🔖

Cite This Article

Ulutaş, E. G., Yüce, E., Süzgen, E. E., Şahin, M. E., Karakuş, M. Ö. (2026). Beyond Accuracy: A Clinically-Oriented Breast Cancer Diagnostic Framework Integrating Conformal Prediction, Counterfactual Explanations, and Multi-Task Learning on FNA Cytological Features. Firat University Journal of Experimental and Computational Engineering, Cilt: 5 Sayı: 2, 552–573. https://doi.org/10.62520/fujece.1940340

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (30)

  1. H. Sung et al., “Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries,” CA Cancer J. Clin., vol. 71, no. 3, pp. 209–249, May 2021.
  2. W. H. Wolberg, W. N. Street, D. M. Heisey, and O. L. Mangasarian, “Computerized Breast Cancer Diagnosis and Prognosis From Fine-Needle Aspirates,” Arch. Surg., vol. 130, no. 5, pp. 511–516, May 1995.
  3. W. Wolberg, O. Mangasarian, N. Street, and W. Street, “Breast Cancer Wisconsin (Diagnostic) - UCI Machine Learning Repository.” Accessed: May 17, 2026. [Online]. Available: https://archive.ics.uci.edu/dataset/17/breast+cancer+wisconsin+diagnostic
  4. J. Owotogbe, E. Oyekanmi, S. E. Adepoju, and A. E. Akinsunmade, “Machine learning and deep learning for breast cancer: A decade systematic review of detection, classification, prognosis, and explainability,” Inform. Med. Unlocked, vol. 63, p. 101756, Jun. 2026.
  5. N. Thakur, P. Kumar, and A. Kumar, “A systematic review of machine and deep learning techniques for the identification and classification of breast cancer through medical image modalities,” Multimed. Tools Appl., vol. 83, no. 12, pp. 35849–35942, Sep. 2023.
  6. S. Hussain, Y. Lafarga-Osuna, M. Ali, U. Naseem, M. Ahmed, and J. G. Tamez-Peña, “Deep learning, radiomics and radiogenomics applications in the digital breast tomosynthesis: a systematic review,” BMC Bioinformatics, vol. 24, no. 1, p. 401, Oct. 2023.
  7. S. Sudarsa and R. P. K. Reddy, “Systematic Review on Breast Cancer Prediction and Classification by using Machine Learning and Deep Learning Methods,” in Proc. 8th Int. Conf. I-SMAC (IoT Social, Mobile, Analytics Cloud) (I-SMAC), 2024, pp. 2049–2059.
  8. Charu and K. Gupta, “Systematic Review of Contemporary Breast Cancer Detection Techniques Using Machine Learning,” in Proc. Int. Conf. Contemp. Comput. Informatics (IC3I), 2024, pp. 338–344.
  9. J. Zhang, Q. Wu, P. Lei, X. Zhu, and B. Li, “Diagnostic accuracy of machine learning-based magnetic resonance imaging models in breast cancer classification: a systematic review and meta-analysis,” World J. Surg. Oncol., vol. 23, no. 1, p. 231, Jun. 2025.
  10. K. A. Abdullah, S. Marziali, M. Nanaa, L. Escudero Sánchez, N. R. Payne, and F. J. Gilbert, “Deep learning-based breast cancer diagnosis in breast MRI: systematic review and meta-analysis,” Eur. Radiol., vol. 35, no. 8, pp. 4474–4489, Feb. 2025.
  11. J. A. Silveira, A. R. da Silva, and M. Z. T. de Lima, “Harnessing artificial intelligence for predicting breast cancer recurrence: a systematic review of clinical and imaging data,” Discov. Oncol., vol. 16, no. 1, p. 135, Feb. 2025.
  12. A. Ghavidel and P. Pazos, “Machine learning (ML) techniques to predict breast cancer in imbalanced datasets: a systematic review,” J. Cancer Surviv., vol. 19, no. 1, pp. 270–294, Feb. 2025.
  13. J. Majidpour and H. Beitollahi, “A Comprehensive Examination of Machine Learning and Deep Learning Approaches for Breast Cancer Detection, Classification, Segmentation, Augmentation, and Feature Selection,” Arch. Comput. Methods Eng., vol. 33, no. 2, pp. 1913–1944, Sep. 2025.
  14. J. Majidpour, H. A. Ahmed, M. H. Ahmed, S. I. Jalal, and H. Arabi, “Applications of GAN Models in Breast Cancer Detection: A Comprehensive Review,” Arch. Comput. Methods Eng., vol. 33, no. 1, pp. 859–915, Aug. 2025.
  15. M. F. Akay, “Support vector machines combined with feature selection for breast cancer diagnosis,” Expert Syst. Appl., vol. 36, no. 2, pp. 3240–3247, Mar. 2009.
  16. G. I. Salama, M. B. Abdelhalim, and M. A. Zeid, “Breast cancer diagnosis on three datasets using multi-classifiers,” Int. J. Comput. Inf. Technol., vol. 1, no. 1, pp. 36–43, Sep. 2012.
  17. O. AlOmair, O. M. Zakaria, M. Alabdullatif, and Z. Khurshid, “Optimized KNN with domain-informed features and LIME explainability for improved breast cancer classification,” BMC Med. Inform. Decis. Mak., vol. 26, no. 1, p. 138, Mar. 2026.
  18. S. Kapoor and A. Narayanan, “Leakage and the reproducibility crisis in machine-learning-based science,” Patterns, vol. 4, no. 9, p. 100804, Sep. 2023.
  19. V. Vovk, A. Gammerman, and G. Shafer, “Algorithmic Learning in a Random World,” in Algorithmic Learning in a Random World. New York, NY, USA: Springer, 2005, pp. 189–221.
  20. A. N. Angelopoulos and S. Bates, “A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantification,” Jul. 2021. Accessed: May 17, 2026. [Online]. Available: https://arxiv.org/pdf/2107.07511
  21. S. M. Lundberg and S.-I. Lee, “A Unified Approach to Interpreting Model Predictions,” Adv. Neural Inf. Process. Syst., vol. 30, 2017. Accessed: May 17, 2026. [Online]. Available: https://github.com/slundberg/shap
  22. A. J. Vickers and E. B. Elkin, “Decision curve analysis: A novel method for evaluating prediction models,” Med. Decis. Making, vol. 26, no. 6, pp. 565–574, Nov. 2006.
  23. H. Olsson et al., “Estimating diagnostic uncertainty in artificial intelligence assisted pathology using conformal prediction,” Nat. Commun., vol. 13, no. 1, p. 7761, Dec. 2022.
  24. A. P. Sreenivasan et al., “Conformal prediction enables disease course prediction and allows individualized diagnostic uncertainty in multiple sclerosis,” npj Digit. Med., vol. 8, no. 1, p. 224, Apr. 2025.
  25. S. Wachter, B. Mittelstadt, and C. Russell, “Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPR,” SSRN Electron. J., Nov. 2017.
  26. R. K. Mothilal, A. Sharma, and C. Tan, “Explaining machine learning classifiers through diverse counterfactual explanations,” in Proc. Conf. Fairness, Accountability, Transparency (FAT)*, 2020, pp. 607–617.
  27. C. Chen et al., “Machine learning algorithms for individualized prediction of prognosis in breast cancer liver metastases and the prognostic impact of primary tumor surgery: a multicenter study,” Front. Endocrinol., vol. 16, p. 1656191, Oct. 2025.
  28. W. Wolberg, W. Street, and O. Mangasarian, “Breast Cancer Wisconsin (Prognostic) - UCI Machine Learning Repository.” Accessed: May 17, 2026. [Online]. Available: https://archive.ics.uci.edu/dataset/16/breast+cancer+wisconsin+prognostic
  29. A. Niculescu-Mizil and R. Caruana, “Predicting good probabilities with supervised learning,” in Proc. 22nd Int. Conf. Mach. Learn. (ICML), 2005, pp. 625–632.
  30. Z. Su et al., “Computational Pathology for Accurate Prediction of Breast Cancer Recurrence: Development and Validation of a Deep Learning–Based Tool,” Mod. Pathol., vol. 38, no. 12, p. 100847, Dec. 2025.