An artificial intelligence-based model exploiting H&E images to predict recurrence in negative sentinel lymph-node melanoma patients.
Artificial intelligence
Digital pathology
Melanoma
Recurrence risk prediction
Journal
Journal of translational medicine
ISSN: 1479-5876
Titre abrégé: J Transl Med
Pays: England
ID NLM: 101190741
Informations de publication
Date de publication:
12 Sep 2024
12 Sep 2024
Historique:
received:
07
03
2024
accepted:
18
08
2024
medline:
13
9
2024
pubmed:
13
9
2024
entrez:
12
9
2024
Statut:
epublish
Résumé
Risk stratification and treatment benefit prediction models are urgent to improve negative sentinel lymph node (SLN-) melanoma patient selection, thus avoiding costly and toxic treatments in patients at low risk of recurrence. To this end, the application of artificial intelligence (AI) could help clinicians to better calculate the recurrence risk and choose whether to perform adjuvant therapy. We made use of AI to predict recurrence-free status (RFS) within 2-years from diagnosis in 94 SLN- melanoma patients. In detail, we detected quantitative imaging information from H&E slides of a cohort of 71 SLN- melanoma patients, who registered at Istituto Tumori "Giovanni Paolo II" in Bari, Italy (investigational cohort, IC). For each slide, two expert pathologists firstly annotated two Regions of Interest (ROIs) containing tumor cells alone (TUMOR ROI) or with infiltrating cells (TUMOR + INF ROI). In correspondence of the two kinds of ROIs, two AI-based models were developed to extract information directly from the tiles in which each ROI was automatically divided. This information was then used to predict RFS. Performances of the models were computed according to a 5-fold cross validation scheme. We further validated the prediction power of the two models on an independent external validation cohort of 23 SLN- melanoma patients (validation cohort, VC). The TUMOR ROIs have revealed more informative than the TUMOR + INF ROIs. An Area Under the Curve (AUC) value of 79.1% and 62.3%, a sensitivity value of 81.2% and 76.9%, a specificity value of 70.0% and 43.3%, an accuracy value of 73.2% and 53.4%, were achieved on the TUMOR and TUMOR + INF ROIs extracted for the IC cohort, respectively. An AUC value of 76.5% and 65.2%, a sensitivity value of 66.7% and 41.6%, a specificity value of 70.0% and 55.9%, an accuracy value of 70.0% and 56.5%, were achieved on the TUMOR and TUMOR + INF ROIs extracted for the VC cohort, respectively. Our approach represents a first effort to develop a non-invasive prognostic method to better define the recurrence risk and improve the management of SLN- melanoma patients.
Sections du résumé
BACKGROUND
BACKGROUND
Risk stratification and treatment benefit prediction models are urgent to improve negative sentinel lymph node (SLN-) melanoma patient selection, thus avoiding costly and toxic treatments in patients at low risk of recurrence. To this end, the application of artificial intelligence (AI) could help clinicians to better calculate the recurrence risk and choose whether to perform adjuvant therapy.
METHODS
METHODS
We made use of AI to predict recurrence-free status (RFS) within 2-years from diagnosis in 94 SLN- melanoma patients. In detail, we detected quantitative imaging information from H&E slides of a cohort of 71 SLN- melanoma patients, who registered at Istituto Tumori "Giovanni Paolo II" in Bari, Italy (investigational cohort, IC). For each slide, two expert pathologists firstly annotated two Regions of Interest (ROIs) containing tumor cells alone (TUMOR ROI) or with infiltrating cells (TUMOR + INF ROI). In correspondence of the two kinds of ROIs, two AI-based models were developed to extract information directly from the tiles in which each ROI was automatically divided. This information was then used to predict RFS. Performances of the models were computed according to a 5-fold cross validation scheme. We further validated the prediction power of the two models on an independent external validation cohort of 23 SLN- melanoma patients (validation cohort, VC).
RESULTS
RESULTS
The TUMOR ROIs have revealed more informative than the TUMOR + INF ROIs. An Area Under the Curve (AUC) value of 79.1% and 62.3%, a sensitivity value of 81.2% and 76.9%, a specificity value of 70.0% and 43.3%, an accuracy value of 73.2% and 53.4%, were achieved on the TUMOR and TUMOR + INF ROIs extracted for the IC cohort, respectively. An AUC value of 76.5% and 65.2%, a sensitivity value of 66.7% and 41.6%, a specificity value of 70.0% and 55.9%, an accuracy value of 70.0% and 56.5%, were achieved on the TUMOR and TUMOR + INF ROIs extracted for the VC cohort, respectively.
CONCLUSIONS
CONCLUSIONS
Our approach represents a first effort to develop a non-invasive prognostic method to better define the recurrence risk and improve the management of SLN- melanoma patients.
Identifiants
pubmed: 39267101
doi: 10.1186/s12967-024-05629-2
pii: 10.1186/s12967-024-05629-2
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
838Subventions
Organisme : Ministero della Salute
ID : RC 2024
Informations de copyright
© 2024. The Author(s).
Références
Ali Z, Yousaf N, Larkin J. Melanoma epidemiology, biology and prognosis. Eur J Cancer Suppl. 2013;11:81–91. https://doi.org/10.1016/j.ejcsup.2013.07.012 .
doi: 10.1016/j.ejcsup.2013.07.012
https://www.cancer.org/cancer/melanoma-skin-cancer/about/key-statistics.html#:~:text=Cancer%20of%20the%20skin%20is,majority%20of%20skin%20cancer%20deaths .
Gershenwald JE, Thompson W, Mansfield PF, et al. Multi-institutional Melanoma Lymphatic Mapping experience: the Prognostic Value of Sentinel Lymph Node Status in 612 stage I or II melanoma patients. J Clin Oncol. 1999;17:976–976. https://doi.org/10.1200/JCO.1999.17.3.976 .
doi: 10.1200/JCO.1999.17.3.976
pubmed: 10071292
Hyams DM, Cook RW, Buzaid AC. Identification of risk in cutaneous melanoma patients: prognostic and predictive markers. J Surg Oncol. 2019;119:175–86. https://doi.org/10.1002/jso.25319 .
doi: 10.1002/jso.25319
pubmed: 30548543
Quaresmini D, Guida M. Neoangiogenesis in Melanoma: an issue in Biology and systemic treatment. Front Immunol. 2020;11. https://doi.org/10.3389/fimmu.2020.584903 .
Balch CM, Gershenwald JE, Soong S, et al. Final version of 2009 AJCC melanoma staging and classification. J Clin Oncol. 2009;27:6199–206. https://doi.org/10.1200/JCO.2009.23.4799 .
doi: 10.1200/JCO.2009.23.4799
pubmed: 19917835
pmcid: 2793035
Yushak M, Mehnert J, Luke J, Poklepovic A. Approaches to high-risk Resected Stage II and III Melanoma. Am Soc Clin Oncol Educ B. 2019;e207–11. https://doi.org/10.1200/EDBK_239283 .
Long GV, Hauschild A, Santinami M, et al. Adjuvant dabrafenib plus Trametinib in Stage III BRAF -Mutated Melanoma. N Engl J Med. 2017;377:1813–23. https://doi.org/10.1056/NEJMoa1708539 .
doi: 10.1056/NEJMoa1708539
pubmed: 28891408
Eggermont AMM, Blank CU, Mandala M, et al. Adjuvant Pembrolizumab versus Placebo in Resected Stage III Melanoma. N Engl J Med. 2018;378:1789–801. https://doi.org/10.1056/NEJMoa1802357 .
doi: 10.1056/NEJMoa1802357
pubmed: 29658430
Weber J, Mandala M, Del Vecchio M, et al. Adjuvant Nivolumab versus Ipilimumab in Resected Stage III or IV Melanoma. N Engl J Med. 2017;377:1824–35. https://doi.org/10.1056/NEJMoa1709030 .
doi: 10.1056/NEJMoa1709030
pubmed: 28891423
Luke JJ, Rutkowski P, Queirolo P, et al. Pembrolizumab versus placebo as adjuvant therapy in completely resected stage IIB or IIC melanoma (KEYNOTE-716): a randomised, double-blind, phase 3 trial. Lancet. 2022;399:1718–29. https://doi.org/10.1016/S0140-6736(22)00562-1 .
doi: 10.1016/S0140-6736(22)00562-1
pubmed: 35367007
Kirkwood JM, Del Vecchio M, Weber J, et al. Adjuvant nivolumab in resected stage IIB/C melanoma: primary results from the randomized, phase 3 CheckMate 76K trial. Nat Med. 2023;29:2835–43. https://doi.org/10.1038/s41591-023-02583-2 .
doi: 10.1038/s41591-023-02583-2
pubmed: 37845511
pmcid: 10667090
Bera K, Schalper KA, Rimm DL, et al. Artificial intelligence in digital pathology — new tools for diagnosis and precision oncology. Nat Rev Clin Oncol. 2019;16:703–15. https://doi.org/10.1038/s41571-019-0252-y .
doi: 10.1038/s41571-019-0252-y
pubmed: 31399699
pmcid: 6880861
Ching T, Himmelstein DS, Beaulieu-Jones BK, et al. Opportunities and obstacles for deep learning in biology and medicine. J R Soc Interface. 2018;15:20170387. https://doi.org/10.1098/rsif.2017.0387 .
doi: 10.1098/rsif.2017.0387
pubmed: 29618526
pmcid: 5938574
Niazi MKK, Parwani AV, Gurcan MN. Digital pathology and artificial intelligence. Lancet Oncol. 2019;20:e253–61. https://doi.org/10.1016/S1470-2045(19)30154-8 .
doi: 10.1016/S1470-2045(19)30154-8
pubmed: 31044723
pmcid: 8711251
Adeuyan O, Gordon ER, Kenchappa D, et al. An update on methods for detection of prognostic and predictive biomarkers in melanoma. Front Cell Dev Biol. 2023;11. https://doi.org/10.3389/fcell.2023.1290696 .
Comes MC, Fucci L, Mele F, et al. A deep learning model based on whole slide images to predict disease-free survival in cutaneous melanoma patients. Sci Rep. 2022;12:20366. https://doi.org/10.1038/s41598-022-24315-1 .
doi: 10.1038/s41598-022-24315-1
pubmed: 36437296
pmcid: 9701687
Hu J, Cui C, Yang W, et al. Using deep learning to predict anti-PD-1 response in melanoma and lung cancer patients from histopathology images. Transl Oncol. 2021;14:100921. https://doi.org/10.1016/j.tranon.2020.100921 .
doi: 10.1016/j.tranon.2020.100921
pubmed: 33129113
Kulkarni PM, Robinson EJ, Pradhan JS, et al. Deep learning based on standard H&E images of primary melanoma tumors identifies patients at risk for visceral recurrence and death. Clin Cancer Res. 2020;26:1126–34. https://doi.org/10.1158/1078-0432.CCR-19-1495 .
doi: 10.1158/1078-0432.CCR-19-1495
pubmed: 31636101
Tjoa E, Guan C. A Survey on Explainable Artificial Intelligence (XAI): toward Medical XAI. IEEE Trans Neural Networks Learn Syst. 2020;32:4793–813. https://doi.org/10.1109/tnnls.2020.3027314 .
doi: 10.1109/tnnls.2020.3027314
Ribeiro MT, Singh S, Guestrin C. (2016) Why Should I Trust You? Explaining the Predictions of Any Classifier. NAACL-HLT 2016–2016 Conf North Am Chapter Assoc Comput Linguist Hum Lang Technol Proc Demonstr Sess 97–101. https://doi.org/10.18653/v1/n16-3020
Mann HB, Whitney DR. On a test of whether one of two Random variables is stochastically larger larger than the other. Ann Math Stat. 1947;18:50–60.
doi: 10.1214/aoms/1177730491
Pandis N. The chi-square test. Am J Orthod Dentofac Orthop. 2016;150:898–9. https://doi.org/10.1016/j.ajodo.2016.08.009 .
doi: 10.1016/j.ajodo.2016.08.009
Pantanowitz L, Farahani N, Parwani A. Whole slide imaging in pathology: advantages, limitations, and emerging perspectives. Pathol Lab Med Int 23. 2015. https://doi.org/10.2147/PLMI.S59826 .
doi: 10.2147/PLMI.S59826
Bankhead P, Loughrey MB, Fernández JA, et al. QuPath: open source software for digital pathology image analysis. Sci Rep. 2017;7:1–7. https://doi.org/10.1038/s41598-017-17204-5 .
doi: 10.1038/s41598-017-17204-5
Macenko M, Niethammer M, Marron JS et al. (2009) A Method For Normalizing Histology Slides For Quantitative Analysis 3 statistics and Operations Research, 4 Lineberger Comprehensive Cancer Center, 5 Renaissance Computing Institute, 6 Pathology and Laboratory Medicine, 7 Dermatology University of nor. IEEE Int Symp Biomed Imaging 1107–10.
Comes MC, Fanizzi A, Bove S, et al. Early prediction of neoadjuvant chemotherapy response by exploiting a transfer learning approach on breast DCE-MRIs. Sci Rep. 2021;11. https://doi.org/10.1038/s41598-021-93592-z .
Chollet F. Xception: deep learning with depthwise separable convolutions. Proc – 30th IEEE Conf Comput Vis Pattern Recognit CVPR 2017. 2017;2017–Janua:1800–7. https://doi.org/10.1109/CVPR.2017.195 .
doi: 10.1109/CVPR.2017.195
Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the Inception Architecture for Computer Vision. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit 2016-Decem. 2016;2818–26. https://doi.org/10.1109/CVPR.2016.308 .
He K. (2015) Deep Residual Learning for Image Recognition ResNet @ ILSVRC & COCO 2015 Competitions. 1–9.
Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. (2017) Densely connected convolutional networks. Proc – 30th IEEE conf Comput Vis Pattern Recognition, CVPR 2017 2017-Janua:2261–9. https://doi.org/10.1109/CVPR.2017.243
Dosovitskiy A, Beyer L, Kolesnikov A et al. (2021) An image is worth 16X16 words: transformers for Image Recognition at Scale. ICLR 2021–9th Int Conf Learn Represent.
Mencattini A, Spalloni A, Casti P, et al. NeuriTES. Monitoring neurite changes through transfer entropy and semantic segmentation in bright-field time-lapse microscopy. Patterns. 2021;2:100261. https://doi.org/10.1016/j.patter.2021.100261 .
doi: 10.1016/j.patter.2021.100261
pubmed: 34179845
pmcid: 8212146
Lin TY, Goyal P, Girshick R, et al. Focal loss for dense object detection. IEEE Trans Pattern Anal Mach Intell. 2020;42:318–27. https://doi.org/10.1109/TPAMI.2018.2858826 .
doi: 10.1109/TPAMI.2018.2858826
pubmed: 30040631
Kingma DP, Ba JL. (2015) Adam: A method for stochastic optimization. 3rd int conf learn represent ICLR 2015 -. Conf Track Proc 1–15.
Amoroso N, Errico R, Bruno S, et al. Hippocampal unified multi-atlas network (HUMAN): protocol and scale validation of a novel segmentation tool. Phys Med Biol. 2015;60:8851–67. https://doi.org/10.1088/0031-9155/60/22/8851 .
doi: 10.1088/0031-9155/60/22/8851
pubmed: 26531765
Garbe C, Keim U, Amaral T, et al. Prognosis of patients with primary Melanoma Stage I and II according to American Joint Committee on Cancer Version 8 validated in two independent cohorts: implications for adjuvant treatment. J Clin Oncol. 2022;40:3741–9. https://doi.org/10.1200/JCO.22.00202 .
doi: 10.1200/JCO.22.00202
pubmed: 35709414
pmcid: 9649277
Bleicher J, Swords DS, Mali ME, et al. Recurrence patterns in patients with stage II melanoma: the evolving role of routine imaging for surveillance. J Surg Oncol. 2020;122:1770–7. https://doi.org/10.1002/jso.26214 .
doi: 10.1002/jso.26214
pubmed: 33098702
Lee R, Mandala M, Long GV, et al. Adjuvant therapy for stage II melanoma: the need for further studies. Eur J Cancer. 2023;189:112914. https://doi.org/10.1016/j.ejca.2023.05.003 .
doi: 10.1016/j.ejca.2023.05.003
pubmed: 37301717
Simiczyjew A, Dratkiewicz E, Mazurkiewicz J, et al. The influence of Tumor Microenvironment on Immune escape of Melanoma. Int J Mol Sci. 2020;21:8359. https://doi.org/10.3390/ijms21218359 .
doi: 10.3390/ijms21218359
pubmed: 33171792
pmcid: 7664679
Liu D, Yang X, Wu X. Tumor Immune Microenvironment characterization identifies prognosis and immunotherapy-related gene signatures in Melanoma. Front Immunol. 2021;12. https://doi.org/10.3389/fimmu.2021.663495 .
Moldoveanu D, Ramsay L, Lajoie M, et al. Spatially mapping the immune landscape of melanoma using imaging mass cytometry. Sci Immunol. 2022;7. https://doi.org/10.1126/sciimmunol.abi5072 .
Serratì S, Di Fonte R, Porcelli L, et al. Circulating extracellular vesicles are monitoring biomarkers of anti-PD1 response and enhancer of tumor progression and immunosuppression in metastatic melanoma. J Exp Clin Cancer Res. 2023;42:251. https://doi.org/10.1186/s13046-023-02808-9 .
doi: 10.1186/s13046-023-02808-9
pubmed: 37759291
pmcid: 10538246