AI Use in Prostate Cancer: Potential Improvements in Treatments and Patient Care

Publication
Article
OncologyONCOLOGY Vol 38, Issue 5
Volume 38
Issue 5
Pages: 208-209

Artificial intelligence use in prostate cancer encompasses 4 main areas including diagnostic imaging, prediction of outcomes, histopathology, and treatment planning.

Artificial intelligence use in prostate cancer encompasses 4 main areas including diagnostic imaging, prediction of outcomes, histopathology, and treatment planning.

Artificial intelligence use in prostate cancer encompasses 4 main areas including diagnostic imaging, prediction of outcomes, histopathology, and treatment planning.

Artificial intelligence (AI) generally describes the concept of computers emulating human intelligence. Used synonymously, machine learning (ML) is considered a subset of AI and describes the field where computers can analyze data and interact with users without explicit coding of each potential possibility. For this review, we will use the term AI, although it will generally refer to concepts of ML. In recent years, AI increasingly has been applied to medicine, oncology, and prostate cancer.1 This review will briefly touch upon 4 areas where AI and prostate cancer have overlapped: AI-driven diagnostic image analysis, AI “prediction” of prostate outcomes based on clinical data, AI prediction using multimodal data including histopathology, and AI definition of tumor and normal tissue for radiation oncology treatment planning. After describing each area, we will give practical examples of application. Finally, we will briefly discuss future applications of AI to prostate cancer.

AI-Driven Diagnostic Image Analysis and Radiomics

Image classification is an area where AI has taken large strides, driven by nonclinical work in computer vision. ML architectures such as convolutional neural networks, and newer network architectures such as transformer-based architectures, have improved the ability of AI to correctly identify elements in photographs. These models are similarly being applied to quantitative characteristics of diagnostic images (known as radiomics) for the purpose of detecting clinically significant disease.2,3 Current published ML models have shown promise but are not able to completely replace radiologist evaluation in real-world situations4; however, they may aid less experienced radiologists in distinguishing between cancerous and noncancerous lesions in prostate MRI scans.5 At the same time, some studies have shown that computer-aided detection (CAD)–assisted mammography may result in reduced sensitivity to non–CAD-identified breast lesions.6 It is possible that overreliance on AI could morph prostate cancer CAD tools from helpful assistants to a second-rate crutch.

AI Predictions of Health Outcomes

ML algorithms have been used to take clinical characteristics and genetic features and predict relevant clinical outcomes such as prostate cancer risk, presence of nodal metastases, response to therapy, and mortality.7-13 Although these ML algorithms may present predictive performance improvements compared with existing nomograms,14 their adoption in routine clinical practice likely will require automated integration into existing health care electronic medical records (EMRs), as well as navigation of regulatory frameworks. Also, obstacles to ML implementation in clinical practice include barriers to real-time data extraction and aggregation from multiple commercial EMR sources and information systems.15 In comparison, a nomogram makes clear the relative contribution of each factor to the intended clinical prediction.


AI Histopathology-Driven Characterization of Prostate Cancer

Evaluation of prostate cancer histopathology is perhaps where the most clinical impact is being made.16 One example where ML tools are helping pathologists categorize prostate cancer is Paige Prostate (Paige AI), a tool for automatically labeling prostate cancer by Gleason score.17 Another example is a multimodal deep learning network that incorporates clinical characteristics as well as features extracted from digitized histopathology to predict outcomes from treatment.18 This model was trained and evaluated in a clinical trials data set to be predictive of androgen deprivation therapy in combination with radiotherapy vs radiotherapy alone.19 Now commercialized as ArteraAI, the multimodal AI test is approved as a clinical diagnostic laboratory test by the Centers for Medicare & Medicaid Services.

AI-Driven Tumor Definition and Treatment Planning for Radiation Therapy

A rapid area of AI expansion is in aiding radiation oncologists in the automated definition of normal tissue and tumor definition. “Contouring” is the general process whereby radiation oncologists delineate organs at risk for radiation toxicity, as well as define radiation treatment targets. The definition of organs at risk and target volumes is traditionally a time-consuming and technically demanding task. AI models are improving the efficiency of this process through automated contouring and are already commercially available.20,21 These models will likely continue to improve in accuracy with recent innovations in ML architecture and multimodal imaging data.22

Once a target volume is defined, AI can improve treatment planning through improvements in efficiency and optimization of dose.23-26 These improvements in efficiency are particularly valuable for online adaptive radiotherapy, where treatment plans are adjusted daily based on time-of-treatment cross-sectional images.27


Future Applications

With the rise of generative AI in day-to-day life, patients will likely use large language model–based tools to obtain cancer treatment information. Physicians may start using AI to perform routine tasks in symptom management and patient-facing interaction.28,29 For example, the System for High-Intensity EvaLuation During Radiation Therapy (SHIELD-RT) study (NCT04277650) found that an ML algorithm accurately identified patients at high risk for needing acute care during radiotherapy. These patients were then able to benefit from random assignment to twice-weekly (vs once-weekly) clinical evaluation.30 It is likely AI will further diffuse into all aspects of health care as a supplemental aid for physicians and patients.31

Conclusion

The innovations seen in the application of AI to prostate cancer care mirror those happening throughout health care and information technology. Breakthroughs in image analysis and computer vision have diffused into the classification of prostate diagnostic imaging, pathology, and prediction of treatment outcomes. Radiation oncology has experienced improvements in practice efficiency due to AI tools. Future applications of AI for prostate cancer likely will include improved patient-facing tools.


References

  1. Chu TN, Wong EY, Ma R, Yang CH, Dalieh IS, Hung AJ. Exploring the use of artificial intelligence in the management of prostate cancer. Curr Urol Rep. 2023;24(5):231-240. doi:10.1007/s11934-023-01149-6
  2. Pachetti E, Colantonio S. 3D-vision-transformer stacking ensemble for assessing prostate cancer aggressiveness from T2w images. Bioengineering (Basel). 2023;10(9):1015. doi:10.3390/bioengineering10091015
  3. Ruan M, Liu Y, Yao K, et al. Development and validation of interpretable machine learning models for clinically significant prostate cancer diagnosis in patients with lesions of PI-RADS v2.1 score ≥3. J Magn Reson Imaging. Published online February 16, 2024. doi:10.1002/jmri.29275
  4. Gresser E, Schachtner B, Stüber AT, et al. Performance variability of radiomics machine learning models for the detection of clinically significant prostate cancer in heterogeneous MRI datasets. Quant Imaging Med Surg. 2022;12(11):4990-5003. doi:10.21037/qims-22-265
  5. Labus S, Altmann MM, Huisman H, et al. A concurrent, deep learning–based computer-aided detection system for prostate multiparametric MRI: a performance study involving experienced and less-experienced radiologists. Eur Radiol. 2023;33(1):64-76. doi:10.1007/s00330-022-08978-y
  6. Lehman CD, Wellman RD, Buist DS, Kerlikowske K, Tosteson AN, Miglioretti DL; Breast Cancer Surveillance Consortium. Diagnostic accuracy of digital screening mammography with and without computer-aided detection. JAMA Intern Med. 2015;175(11):1828-1837. doi:10.1001/jamainternmed.2015.5231
  7. Yin W, Chen G, Li Y, et al. Identification of a 9-gene signature to enhance biochemical recurrence prediction in primary prostate cancer: a benchmarking study using ten machine learning methods and twelve patient cohorts. Cancer Lett. Published online February 22, 2024. doi:10.1016/j.canlet.2024.216739
  8. Dai X, Park JH, Yoo S, et al. Survival analysis of localized prostate cancer with deep learning. Sci Rep. 2022;12(1):17821. doi:10.1038/s41598-022-22118-y
  9. De Bari B, Vallati M, Gatta R, et al. Could machine learning improve the prediction of pelvic nodal status of prostate cancer patients? Preliminary results of a pilot study. Cancer Invest. 2015;33(6):232-240. doi:10.3109/07357907.2015.1024317
  10. Koo KC, Lee KS, Kim S, et al. Long short-term memory artificial neural network model for prediction of prostate cancer survival outcomes according to initial treatment strategy: development of an online decision-making support system. World J Urol. 2020;38(10):2469-2476. doi:10.1007/s00345-020-03080-8
  11. Erho N, Crisan A, Vergara IA, et al. Discovery and validation of a prostate cancer genomic classifier that predicts early metastasis following radical prostatectomy. PLoS One. 2013;8(6):e66855. doi:10.1371/journal.pone.0066855
  12. Sabbagh A, Tilki D, Feng J, et al. Multi-institutional development and external validation of a machine learning model for the prediction of distant metastasis in patients treated by salvage radiotherapy for biochemical failure after radical prostatectomy. Eur Urol Focus. 2024;10(1):66-74. doi:10.1016/j.euf.2023.07.004
  13. Sabbagh A, Washington SL 3rd, Tilki D, et al. Development and external validation of a machine learning model for prediction of lymph node metastasis in patients with prostate cancer. Eur Urol Oncol. 2023;6(5):501-507. doi:10.1016/j.euo.2023.02.006
  14. Tan YG, Fang AHS, Lim JKS, et al. Incorporating artificial intelligence in urology: supervised machine learning algorithms demonstrate comparative advantage over nomograms in predicting biochemical recurrence after prostatectomy. Prostate. 2022;82(3):298-305. doi:10.1002/pros.24272
  15. Hong JC, Eclov NCW, Stephens SJ, Mowery YM, Palta M. Implementation of machine learning in the clinic: challenges and lessons in prospective deployment from the System for High Intensity EvaLuation During Radiation Therapy (SHIELD-RT) randomized controlled study. BMC Bioinformatics. 2022;23(suppl 12):408. doi:10.1186/s12859-022-04940-3
  16. Frewing A, Gibson AB, Robertson R, Urie PM, Corte DD. Don’t fear the artificial intelligence: a systematic review of machine learning for prostate cancer detection in pathology. Arch Pathol Lab Med. Published online August 18, 2023. doi:10.5858/arpa.2022-0460-RA
  17. Raciti P, Sue J, Ceballos R, et al. Novel artificial intelligence system increases the detection of prostate cancer in whole slide images of core needle biopsies. Mod Pathol. 2020;33(10):2058-2066. doi:10.1038/s41379-020-0551-y
  18. Esteva A, Feng J, van der Wal D, et al; NRG Prostate Cancer AI Consortium. Prostate cancer therapy personalization via multi-modal deep learning on randomized phase III clinical trials NPJ Digit Med. 2022;5(1):71. doi:10.1038/s41746-022-00613-w
  19. Spratt DE, Tang S, Sun Y, et al. Artificial intelligence predictive model for hormone therapy use in prostate cancer. Preprint. Res Sq. 2023;rs.3.rs-2790858. Published 2023 Apr 21. doi:10.21203/rs.3.rs-2790858/v1
  20. Sarria GR, Kugel F, Roehner F, et al. Artificial intelligence–based autosegmentation: advantages in delineation, absorbed dose-distribution, and logistics. Adv Radiat Oncol. 2023;9(3):101394. doi:10.1016/j.adro.2023.101394
  21. Baroudi H, Brock KK, Cao W, et al. Automated contouring and planning in radiation therapy: what is ‘clinically acceptable’? Diagnostics (Basel). 2023;13(4):667. doi:10.3390/diagnostics13040667
  22. Li Y, Wu Y, Huang M, Zhang Y, Bai Z. Attention-guided multi-scale learning network for automatic prostate and tumor segmentation on MRI. Comput Biol Med. 2023;165:107374. doi:10.1016/j.compbiomed.2023.107374
  23. Winter JD, Reddy V, Li W, Craig T, Raman S. Impact of technological advances in treatment planning, image guidance, and treatment delivery on target margin design for prostate cancer radiotherapy: an updated review. Br J Radiol. 2024;97(1153):31-40. doi:10.1093/bjr/tqad041
  24. Dickhoff LRM, Scholman RJ, Barten DLJ, et al. Keeping your best options open with AI-based treatment planning in prostate and cervix brachytherapy. Brachytherapy. 2024;23(2):188-198. doi:10.1016/j.brachy.2023.10.005
  25. Kadoya N, Kimura Y, Tozuka R, et al. Evaluation of deep learning–based deliverable VMAT plan generated by prototype software for automated planning for prostate cancer patients. J Radiat Res. 2023;64(5):842-849. doi:10.1093/jrr/rrad058
  26. McIntosh C, Conroy L, Tjong MC, et al. Clinical integration of machine learning for curative-intent radiation treatment of patients with prostate cancer. Nat Med. 2021;27(6):999-1005. doi:10.1038/s41591-021-01359-w
  27. Roberfroid B, Barragán-Montero AM, Dechambre D, Sterpin E, Lee JA, Geets X. Comparison of Ethos template-based planning and AI-based dose prediction: general performance, patient optimality, and limitations. Phys Med. 2023;116:103178. doi:10.1016/j.ejmp.2023.103178
  28. Altamimi I, Altamimi A, Alhumimidi AS, Altamimi A, Temsah MH. Artificial intelligence (AI) chatbots in medicine: a supplement, not a substitute. Cureus. 2023;15(6):e40922. doi:10.7759/cureus.40922
  29. Chen S, Kann BH, Foote MB, et al. Use of artificial intelligence chatbots for cancer treatment information. JAMA Oncol. 2023;9(10):1459-1462. doi:10.1001/jamaoncol.2023.2954
  30. Hong JC, Eclov NCW, Dalal NH, et al. System for high-intensity evaluation during radiation therapy (SHIELD-RT): a prospective randomized study of machine learning–directed clinical evaluations during radiation and chemoradiation. J Clin Oncol. 2020;38(31):3652-3661. doi:10.1200/JCO.20.01688
  31. Alowais SA, Alghamdi SS, Alsuhebany N, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ. 2023;23(1):689. doi:10.1186/s12909-023-04698-z
Related Videos
The use of CT scans may help practices adaptively plan and adjust radiotherapy courses for patients with non–small cell lung cancer.
Patients with NSCLC who have comorbidities or frailty may also be able to receive treatment with fewer toxicities via proton beam radiotherapy.
Terrence T. Sio, MD, MS, emphasizes multidisciplinary collaboration for treating patients with NSCLC who may require more than 1 type of therapy.
The use of proton therapy may offer a more specific depth charge compared with conventional radiation, according to Timothy Chen, MD.
Two women in genitourinary oncology discuss their experiences with figuring out when to begin a family and how to prioritize both work and children.
Over the past few decades, the prostate cancer space has evolved with increased funding for clinical trial creation and enrollment.
Prophylactic cranial irradiation may not be worthwhile for treating patients with extensive-stage small cell lung cancer based on conflicting data, according to Gregory Peter Kalemkerian, MD.
Rohit Gosain, MD; Rahul Gosain, MD; and Rana R. McKay, MD, presenting slides
Rohit Gosain, MD; Rahul Gosain, MD; and Rana R. McKay, MD, presenting slides
Rohit Gosain, MD; Rahul Gosain, MD; and Rana R. McKay, MD, presenting slides
Related Content