Cheng Jiang1, Abhishek Bhattacharya2, Joseph R Linzey3, Rushikesh S Joshi3, Sung Jik Cha4, Sudharsan Srinivasan2, Daniel Alber5, Akhil Kondepudi6, Esteban Urias7, Balaji Pandian2, Wajd N Al-Holou3, Stephen E Sullivan3, B Gregory Thompson3, Jason A Heth3, Christian W Freudiger8, Siri Sahib S Khalsa3, Donato R Pacione9, John G Golfinos9, Sandra Camelo-Piragua10, Daniel A Orringer9,11, Honglak Lee12, Todd C Hollon3. 1. Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, Michigan, USA. 2. School of Medicine, University of Michigan, Ann Arbor, Michigan, USA. 3. Department of Neurosurgery, University of Michigan, Ann Arbor, Michigan, USA. 4. School of Medicine, Western Michigan University, Kalamazoo, Michigan, USA. 5. Division of Applied Mathematics, Brown University, Providence, Rhode Island, USA. 6. College of Literature, Science and the Arts, University of Michigan, Ann Arbor, Michigan, USA. 7. Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan, USA. 8. Invenio Imaging, Inc., Santa Clara, California, USA. 9. Department of Neurosurgery, NYU Langone Health, New York, New York, USA. 10. Department of Pathology, University of Michigan, Ann Arbor, Michigan, USA. 11. Department of Pathology, NYU Langone Health, New York, New York, USA. 12. Department of Computer Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.
Abstract
BACKGROUND: Accurate specimen analysis of skull base tumors is essential for providing personalized surgical treatment strategies. Intraoperative specimen interpretation can be challenging because of the wide range of skull base pathologies and lack of intraoperative pathology resources. OBJECTIVE: To develop an independent and parallel intraoperative workflow that can provide rapid and accurate skull base tumor specimen analysis using label-free optical imaging and artificial intelligence. METHODS: We used a fiber laser-based, label-free, nonconsumptive, high-resolution microscopy method (<60 seconds per 1 × 1 mm2), called stimulated Raman histology (SRH), to image a consecutive, multicenter cohort of patients with skull base tumor. SRH images were then used to train a convolutional neural network model using 3 representation learning strategies: cross-entropy, self-supervised contrastive learning, and supervised contrastive learning. Our trained convolutional neural network models were tested on a held-out, multicenter SRH data set. RESULTS: SRH was able to image the diagnostic features of both benign and malignant skull base tumors. Of the 3 representation learning strategies, supervised contrastive learning most effectively learned the distinctive and diagnostic SRH image features for each of the skull base tumor types. In our multicenter testing set, cross-entropy achieved an overall diagnostic accuracy of 91.5%, self-supervised contrastive learning 83.9%, and supervised contrastive learning 96.6%. Our trained model was able to segment tumor-normal margins and detect regions of microscopic tumor infiltration in meningioma SRH images. CONCLUSION: SRH with trained artificial intelligence models can provide rapid and accurate intraoperative analysis of skull base tumor specimens to inform surgical decision-making.
BACKGROUND: Accurate specimen analysis of skull base tumors is essential for providing personalized surgical treatment strategies. Intraoperative specimen interpretation can be challenging because of the wide range of skull base pathologies and lack of intraoperative pathology resources. OBJECTIVE: To develop an independent and parallel intraoperative workflow that can provide rapid and accurate skull base tumor specimen analysis using label-free optical imaging and artificial intelligence. METHODS: We used a fiber laser-based, label-free, nonconsumptive, high-resolution microscopy method (<60 seconds per 1 × 1 mm2), called stimulated Raman histology (SRH), to image a consecutive, multicenter cohort of patients with skull base tumor. SRH images were then used to train a convolutional neural network model using 3 representation learning strategies: cross-entropy, self-supervised contrastive learning, and supervised contrastive learning. Our trained convolutional neural network models were tested on a held-out, multicenter SRH data set. RESULTS: SRH was able to image the diagnostic features of both benign and malignant skull base tumors. Of the 3 representation learning strategies, supervised contrastive learning most effectively learned the distinctive and diagnostic SRH image features for each of the skull base tumor types. In our multicenter testing set, cross-entropy achieved an overall diagnostic accuracy of 91.5%, self-supervised contrastive learning 83.9%, and supervised contrastive learning 96.6%. Our trained model was able to segment tumor-normal margins and detect regions of microscopic tumor infiltration in meningioma SRH images. CONCLUSION: SRH with trained artificial intelligence models can provide rapid and accurate intraoperative analysis of skull base tumor specimens to inform surgical decision-making.
Authors: Todd Hollon; Sandra I Camelo-Piragua; Erin L McKean; Stephen E Sullivan; Hugh J L Garton Journal: World Neurosurg Date: 2015-08-24 Impact factor: 2.104
Authors: Tiffany Y Chen; Drew F K Williamson; Ming Y Lu; Melissa Zhao; Maha Shady; Jana Lipkova; Faisal Mahmood Journal: Nature Date: 2021-05-05 Impact factor: 49.962
Authors: Andre Esteva; Brett Kuprel; Roberto A Novoa; Justin Ko; Susan M Swetter; Helen M Blau; Sebastian Thrun Journal: Nature Date: 2017-01-25 Impact factor: 49.962
Authors: Ronald M Juarez-Chambi; Carmen Kut; Jose J Rico-Jimenez; Kaisorn L Chaichana; Jiefeng Xi; Daniel U Campos-Delgado; Fausto J Rodriguez; Alfredo Quinones-Hinojosa; Xingde Li; Javier A Jo Journal: Clin Cancer Res Date: 2019-07-17 Impact factor: 12.531
Authors: Christian W Freudiger; Wei Min; Brian G Saar; Sijia Lu; Gary R Holtom; Chengwei He; Jason C Tsai; Jing X Kang; X Sunney Xie Journal: Science Date: 2008-12-19 Impact factor: 47.728
Authors: David B Altshuler; Chris A Andrews; Hemant A Parmar; Stephen E Sullivan; Jonathan D Trobe Journal: J Neuroophthalmol Date: 2021-12-01 Impact factor: 3.042
Authors: Daniel A Orringer; Balaji Pandian; Yashar S Niknafs; Todd C Hollon; Julianne Boyle; Spencer Lewis; Mia Garrard; Shawn L Hervey-Jumper; Hugh J L Garton; Cormac O Maher; Jason A Heth; Oren Sagher; D Andrew Wilkinson; Matija Snuderl; Sriram Venneti; Shakti H Ramkissoon; Kathryn A McFadden; Amanda Fisher-Hubbard; Andrew P Lieberman; Timothy D Johnson; X Sunney Xie; Jay K Trautman; Christian W Freudiger; Sandra Camelo-Piragua Journal: Nat Biomed Eng Date: 2017-02-06 Impact factor: 25.671
Authors: Todd C Hollon; Balaji Pandian; Esteban Urias; Akshay V Save; Arjun R Adapa; Sudharsan Srinivasan; Neil K Jairath; Zia Farooq; Tamara Marie; Wajd N Al-Holou; Karen Eddy; Jason A Heth; Siri Sahib S Khalsa; Kyle Conway; Oren Sagher; Jeffrey N Bruce; Peter Canoll; Christian W Freudiger; Sandra Camelo-Piragua; Honglak Lee; Daniel A Orringer Journal: Neuro Oncol Date: 2021-01-30 Impact factor: 12.300
Authors: David Reinecke; Niklas von Spreckelsen; Christian Mawrin; Adrian Ion-Margineanu; Gina Fürtjes; Stephanie T Jünger; Florian Khalid; Christian W Freudiger; Marco Timmer; Maximilian I Ruge; Roland Goldbrunner; Volker Neuschmelting Journal: Acta Neuropathol Commun Date: 2022-08-06 Impact factor: 7.578