From Pixels to Prognosis: Is Artificial Intelligence Ready for Primetime in Management of Head and Neck Cutaneous Cancers?

Alison Hawkins¹, BA and Dr. Ameya Asarkar^1,2, MD

¹ Louisiana State University Health Shreveport School of Medicine, Shreveport, LA

² Department of Otolaryngology-Head and Neck Surgery, LSU Health Shreveport, LA

As head and neck surgeons, we are constantly seeking new methods and tools to best care for our patients. Among these innovations, artificial intelligence (AI) is transforming the landscape of cutaneous malignancy diagnosis and treatment. AI-driven tools now assist in early recognition, histopathologic interpretation, prognostic modeling, and treatment planning, thus offering us new ways to improve patient outcomes. As AI continues to gain popularity, it is essential that we are familiar with both the capabilities and limitations of its applications for managing skin cancer.

Early Detection and Diagnosis

AI can aid early detection of cutaneous malignancies through its machine learning and deep learning techniques by registering data from clinical and dermoscopic images and then classifying the contents of the images by lesion size, shape, borders, and other possible characteristic variabilities¹. From the deep learning AI algorithm, Convolutional Neural Networks (CNNs) have shown to be a promising option in clustering and classifying images with the highest level of accuracy.¹ They work by taking pre-processed images from a database (for example, the International Skin Imaging Collaboration, or ISIC, database) and reviewing them in slices or layers, to identify patterns^1–3. However, image quality is an important consideration when using these technologies, and dermoscopic images with better pixelation seem to provide better accuracy than a traditional camera image.⁴

• In 2017, Esteva et. al. first proposed that CNNs could expertly distinguish benign vs malignant skin lesions.⁵
• A 2024 systematic review found that clinicians benefit from adjunct AI assistance in skin cancer diagnosis.⁶
• The lack of standardization of camera images, especially those taken by patients, is a limitation to the utility of AI image assessment as a diagnostic tool. ⁷

Being able to snap a photo of lesions in real-time with a tailored algorithm allows for potentially expedited detection and can decrease time to diagnosis when used as an adjunct tool for clinicians. 

Histopathologic Interpretation

Digital pathology and the application of AI models allow pathologists to compare images to large datasets in order to better identify patterns and make predictions. Multiple studies assessing histopathologic interpretation by AI algorithms have shown high levels of accuracy.^8,9 Once again, studies emphasize that AI should be used as a supplemental tool for clinical decision-making as opposed to a stand-alone resource, as the accuracy of models improves under clinician surveillance.¹⁰

Prognostic and Treatment Modeling

The use of AI models extends beyond diagnosis of skin cancers to the development of prognostic and treatment models. These algorithms can take into account predictors such as TNM staging and ulceration to then assess outcomes such as likelihood of recurrence.

• One model proposed by Stanford was found to have a 90% sensitivity for predicting 5-year melanoma recurrence, highlighting the promising potential of AI in skin cancer prognosis.¹¹
• Additional research assessing AI in skin cancer prognosis also found models to be helpful in determining both short-term prognosis and likelihood of metastasis based on both patient and tumor characteristics.^12,13
• Other studies have shown that AI demonstrates a high level of accuracy when predicting immunotherapy response in melanoma.^14–16

What about the cost?

While cost is an important consideration when looking to incorporate AI in our practices, there is limited data outlining the cost of AI operations.

• One study analyzed insurance data to estimate out-of-pocket costs from the payer’s perspective for cases with comparable outcomes managed by dermatologists with and without AI assistance. The results showed a slightly lower mean cost with AI ($750) compared to without AI ($759).¹⁷
• Another study found that AI increased costs for dermatologists when made accessible to patients in the form of an app, as more patients claimed to have malignancies and required further costly workup with referrals and false positives.^18,19

With little available research outlining the cost of AI models for cutaneous lesions, it is difficult to predict the general cost of these services for hospitals and physicians. While decreasing time to diagnosis may expedite treatment and thus decrease overall costs, overutilization of resources by patients may prove to be more costly.

Who is using it more? Head and Neck Surgeons vs Dermatologists

There is little quantitative data directly comparing the adoption of AI technologies by dermatologists and head and neck surgeons. Based on current literature, it appears that the two fields might utilize the services in different ways. Dermatologists seem to be utilizing AI for more routine lesion screening, whereas head and neck surgeons have been focusing more on advanced cancer staging, surgical planning, and treatment. More specifically, head and neck surgeons appear to be using AI for reconstruction and graft planning for cancers of various regions of the head and neck as opposed to just cutaneous malignancies.²⁰

Is patient data protected?

One important consideration when utilizing AI to analyze patient images is maintaining patient privacy.  Studies have shown that patients who use smartphone applications to send pictures of suspicious skin lesions indicated they would not have seen a specialist without these programs, underscoring a positive effect on patient engagement with their healthcare.⁴ However, some patients have been found to view AI negatively, with significant concerns about their privacy.^21,22

• AI systems are susceptible to cyber-attacks and add another online space for potential breach of patient data.
• Many systems focus on patient de-identification as opposed to protecting patient privacy head-on²³.
• Privacy measures can be costly to implement, posing another limitation to the use of AI in practice.²⁴

 Patient privacy is an ongoing challenge, and although models have been proposed, full preservation of privacy has not yet been achieved.

Conclusion:

While artificial intelligence provides an exciting new modality for managing and treating our patients, many uncertainties remain regarding its accessibility and utility. It is a compelling tool in the early work-up of skin cancers but has potential limitations including cost and patient privacy. Additional research is needed to specifically assess how AI is being used by head and neck surgeons for cutaneous lesions in comparison to dermatologists. Ultimately, as we continue to explore AI in our practices, it is essential that we integrate these tools thoughtfully and keep optimal patient care at the forefront of our decision-making.

References

1. Bechelli S, Delhommelle J. Machine Learning and Deep Learning Algorithms for Skin Cancer Classification from Dermoscopic Images. Bioengineering. 2022;9(3):97. doi:10.3390/bioengineering9030097
2. Lee JRH, Pavlova M, Famouri M, Wong A. Cancer-Net SCa: tailored deep neural network designs for detection of skin cancer from dermoscopy images. BMC Med Imaging. 2022;22:143. doi:10.1186/s12880-022-00871-w
3. Parvathaneni A, Taylor MM, Black TA, Nelson KC. International Survey on Dermoscopic Image Management: ISIC Data on Capture, Storage, and AI Integration in Dermatology. Dermatol Pract Concept. 2025;15(2):4896. doi:10.5826/dpc.1502a4896
4. Sun MD, Kentley J, Mehta P, Dusza S, Halpern AC, Rotemberg V. Accuracy of commercially available smartphone applications for the detection of melanoma. Br J Dermatol. 2022;186(4):744-746. doi:10.1111/bjd.20903
5. Esteva A, Kuprel B, Novoa RA, et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017;542(7639):115-118. doi:10.1038/nature21056
6. Krakowski I, Kim J, Cai ZR, et al. Human-AI interaction in skin cancer diagnosis: a systematic review and meta-analysis. Npj Digit Med. 2024;7(1):78. doi:10.1038/s41746-024-01031-w
7. Brancaccio G, Balato A, Malvehy J, Puig S, Argenziano G, Kittler H. Artificial Intelligence in Skin Cancer Diagnosis: A Reality Check. J Invest Dermatol. 2024;144(3):492-499. doi:10.1016/j.jid.2023.10.004
8. Mosquera-Zamudio A, Launet L, Tabatabaei Z, et al. Deep Learning for Skin Melanocytic Tumors in Whole-Slide Images: A Systematic Review. Cancers. 2022;15(1):42. doi:10.3390/cancers15010042
9. Mukhopadhyay S, Feldman MD, Abels E, et al. Whole Slide Imaging Versus Microscopy for Primary Diagnosis in Surgical Pathology. Am J Surg Pathol. 2018;42(1):39-52. doi:10.1097/PAS.0000000000000948
10. Hart SN, Flotte W, Norgan AP, et al. Classification of Melanocytic Lesions in Selected and Whole-Slide Images via Convolutional Neural Networks. J Pathol Inform. 2019;10:5. doi:10.4103/jpi.jpi_32_18
11. Xiang J, Wang X, Zhang X, et al. A vision–language foundation model for precision oncology. Nature. 2025;638(8051):769-778. doi:10.1038/s41586-024-08378-w
12. Xu C, Yu X, Ding Z, et al. Artificial intelligence-assisted metastasis and prognosis model for patients with nodular melanoma. PLOS ONE. 2024;19(8):e0305468. doi:10.1371/journal.pone.0305468
13. Cozzolino C, Buja A, Rugge M, et al. Machine learning to predict overall short-term mortality in cutaneous melanoma. Discov Oncol. 2023;14:13. doi:10.1007/s12672-023-00622-5
14. Li J, Dan K, Ai J. Machine learning in the prediction of immunotherapy response and prognosis of melanoma: a systematic review and meta-analysis. Front Immunol. 2024;15. doi:10.3389/fimmu.2024.1281940
15. Tabari A, Cox M, D’Amore B, et al. Machine Learning Improves the Prediction of Responses to Immune Checkpoint Inhibitors in Metastatic Melanoma. Cancers. 2023;15(10):2700. doi:10.3390/cancers15102700
16. Gschwind A, Ossowski S. AI Model for Predicting Anti-PD1 Response in Melanoma Using Multi-Omics Biomarkers. Cancers. 2025;17(5):714. doi:10.3390/cancers17050714
17. Gomez Rossi J, Rojas-Perilla N, Krois J, Schwendicke F. Cost-effectiveness of Artificial Intelligence as a Decision-Support System Applied to the Detection and Grading of Melanoma, Dental Caries, and Diabetic Retinopathy. JAMA Netw Open. 2022;5(3):e220269. doi:10.1001/jamanetworkopen.2022.0269
18. Venkatesh KP, Raza M, Kvedar J. AI-based skin cancer detection: the balance between access and overutilization. NPJ Digit Med. 2023;6:147. doi:10.1038/s41746-023-00900-0
19. Smak Gregoor AM, Sangers TE, Bakker LJ, et al. An artificial intelligence based app for skin cancer detection evaluated in a population based setting. NPJ Digit Med. 2023;6(1):90. doi:10.1038/s41746-023-00831-w
20. Pham TD, Teh MT, Chatzopoulou D, Holmes S, Coulthard P. Artificial Intelligence in Head and Neck Cancer: Innovations, Applications, and Future Directions. Curr Oncol. 2024;31(9):5255-5290. doi:10.3390/curroncol31090389
21. Tai K, Zhao R, Rameau A. Patient-Centered Equitable and Safe Artificial Intelligence in Otolaryngology – Head and Neck Surgery. Otolaryngol–Head Neck Surg Off J Am Acad Otolaryngol-Head Neck Surg. 2024;171(4):1232-1235. doi:10.1002/ohn.881
22. Nelson CA, Pérez-Chada LM, Creadore A, et al. Patient Perspectives on the Use of Artificial Intelligence for Skin Cancer Screening: A Qualitative Study. JAMA Dermatol. 2020;156(5):501-512. doi:10.1001/jamadermatol.2019.5014
23. Ziller A, Mueller TT, Stieger S, et al. Reconciling privacy and accuracy in AI for medical imaging. Nat Mach Intell. 2024;6(7):764-774. doi:10.1038/s42256-024-00858-y
24. Reconciling privacy and accuracy in AI for medical imaging | Nature Machine Intelligence. Accessed October 29, 2025. https://www.nature.com/articles/s42256-024-00858-y