Predicting anatomic pathology turnaround time using machine learning models

Abstract

Turnaround Time (TAT) is a critical quality indicator in anatomic pathology laboratories, yet its management remains a challenge, particularly in resource-limited public hospitals in Thailand. While recent work shows the potential of machine learning (ML) for operational intelligence, studies focusing on TAT prediction in this specific context are scarce. This study aimed to develop and evaluate ML models for predicting the TAT of anatomic pathology specimens at Sakon Nakhon Hospital and to identify the key factors influencing TAT. A retrospective dataset of 37,515 cleaned pathology records was utilized. Feature engineering was performed to create variables reflecting workload (e.g., daily case count) and temporality (e.g., week of the year). Several models, including Random Forest and LightGBM, were developed and compared, with Mean Absolute Error (MAE) as the primary performance metric. The optimized Random Forest model demonstrated the highest performance, achieving an MAE of 1.85 days and an R-squared (R²) of 0.91 on the test set. Feature importance analysis revealed that the assigned pathologist, week of the year, and daily case count were the most significant predictors, highlighting the impact of personnel, temporal, and workload factors. Machine learning models, particularly Random Forest, show high potential for accurately predicting TAT. The findings can be leveraged to develop a decision support tool for proactive laboratory management, aiming to enhance operational efficiency and improve the quality of patient care in resource-constrained healthcare environments.

References

G. Hanna et al., “Recommendations for performance evaluation of machine learning in pathology: A concept paper from the college of american pathologists,” Arch Pathol Lab Med, vol. 148, no. 10, pp. e335–e361, Oct. 2024.

Frewing, A. B. Gibson, R. Robertson, P. M. Urie, and D. Della Corte, “A systematic review of machine learning for prostate cancer detection in pathology,” Arch Pathol Lab Med, vol. 148, no. 5, pp. 603–612, May 2024.

Breil, F. Fritz, V. Thiemann, and M. Dugas, “Mapping turnaround times (TAT) to a generic timeline: A systematic review of TAT definitions in clinical domains,” BMC Med Inform Decis Mak, vol. 11, no. 1, p. 34, May 2011.

E. Volmar, M. O. Idowu, R. J. Souers, D. S. Karcher, and R. E. Nakhleh, “Turnaround time for large or complex specimens in surgical pathology: A college of american pathologists Q-probes study of 56 institutions,” Arch Pathol Lab Med, vol. 139, no. 2, pp. 171–177, Feb. 2015.

S. B. D. D. C. Martins, R. A. C. Leite, and G. C. D. S. Lamb, “Anxiety and depression in the preoperative period of cardiac surgery,” Revista Brasileira de Cirurgia Cardiovascular, vol. 31, pp. 243–251, Jun. 2016.

O. S. Júnior, R. de C. S. A. M. de Deus, and P. S. F. de Sousa, “A systematic review of the literature on the use of lean healthcare in the radiology sector,” Journal of Health Management, vol. 26, no. 2, pp. 297–313, Apr. 2024.

พ. วรรณไกรโรจน์, “ภาพรวมของมาตรฐานห้องปฏิบัติ การพยาธิวิทยากายวิภาค,” ราชวิทยาลัยพยาธิแพทย์แห่ง ประเทศไทย, 2560.

S. Ahluwalia et al., “The need for computational pathology in resource-limited settings,” JCO Global Oncology, no. 8, p. e2200044, Aug. 2022.

Acs, “Telepathology for resource-limited settings,” J. Am. Soc. Cytopathol., vol. 11, no. 5, pp. 291–304, Sep. 2022.

S. McClintock, “Informatics and digital pathology,” Pathobiology, vol. 83, no. 2–3, pp. 73–82, 2016.

E. D’Alfonso and A. A. T. M. Reinders, “Artificial intelligence for digital pathology: a review of the literature and overview of a single-institution experience,” Seminars in Diagnostic Pathology, vol. 40, no. 7, pp. 524–532, Nov. 2023.

H. Harrison et al., “Introduction to artificial intelligence and machine learning for pathology,” Arch Pathol Lab Med, vol. 145, no. 10, pp. 1228–1254, Oct. 2021.

Z. Alom et al., “A state-of-the-art survey on deep learning theory and architectures,” Electronics, vol. 8, no. 3, p. 292, Mar. 2019.

van der Byl et al., “Artificial intelligence in diagnostic pathology: a review of the literature and overview of a single-institution experience,” Seminars in Diagnostic Pathology, vol. 40, no. 7, pp. 524–532, Nov. 2023.

B. Nassif, O. Al-Dahidi, S. Al-Ali, and M. A. Talib, “The battle of machine learning and deep learning in healthcare: A survey,” Journal of Big Data, vol. 11, no. 1, p. 95, Dec. 2024.

M. Abdul Latiff et al., “Voice pathology detection using machine learning algorithms based on different voice databases,” Results in Engineering, vol. 25, p. 103937, Mar. 2025.

M. T. Gebru, J. A. M. da Silva, A. M. da S. A. Filho, J. M. da S. A. Junior, and S. R. L. da Silva, “Turnaround time analysis using data mining from a laboratory information system,” The Journal of Health Care Organization, Provision, and Financing, vol. 62, Jan. 2025.

Karaca, “A comprehensive review of the evolution of deep learning: a recent survey,” Multimedia Tools and Applications, vol. 83, no. 17, pp. 50921–50974, Jun. 2024.

Reddy et al., “Predicting chronic kidney disease progression using small pathology datasets and explainable machine learning models,” Comput Methods Programs Biomed Updat, vol. 6, p. 100160, 2024.

van der Byl, D. van der Byl, P. Cooper, A. van der Merwe, J. Schneider, and J. M. Warwick, “The operationalisation of artificial intelligence in anatomical pathology,” J Clin Pathol, vol. 76, no. 4, pp. 225–231, Apr. 2023