{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"bioRxiv"}],"indexed":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T17:41:57Z","timestamp":1785433317580,"version":"3.56.0"},"posted":{"date-parts":[[2026,3,13]]},"group-title":"Bioengineering","reference-count":19,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,3,13]],"date-time":"2026-03-13T00:00:00Z","timestamp":1773360000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"award":["R01EY021200"],"award-info":[{"award-number":["R01EY021200"]}],"id":[{"id":"https:\/\/ror.org\/03wkg3b53","id-type":"ROR","asserted-by":"publisher"}]},{"award":["P50DA037844"],"award-info":[{"award-number":["P50DA037844"]}],"id":[{"id":"https:\/\/ror.org\/00fq5cm18","id-type":"ROR","asserted-by":"publisher"}]},{"award":["Challenge Grant"],"award-info":[{"award-number":["Challenge Grant"]}],"id":[{"id":"https:\/\/ror.org\/04drjs621","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,3,13]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                <jats:sec>\n                  <jats:title>Purpose<\/jats:title>\n                  <jats:p>Machine learning approaches for automated quantification of optic nerve histology have emerged as potential tools for objective assessment of axonal injury in experimental glaucoma models. However, the generalizability of these models to independent datasets remains unclear. Guided by a scoping review of the literature, this study performed independent validation testing of publicly available models on a novel rat optic nerve dataset to assess their generalizability.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We conducted a scoping review following PRISMA-ScR guidelines. PubMed, EMBASE, Scopus, and Cochrane CENTRAL were searched from 2000 through 2025. Two reviewers independently screened records and extracted data on model characteristics and performance metrics. Additionally, we performed independent validation of three models (AxoNet, AxonDeepSeg, AxoNet 2.0) on a novel rat optic nerve dataset comprising 57 images with 9,514 manually annotated axons. Because AxonDeep is not publicly available, we instead evaluated AxonDeepSeg, a separate publicly available deep learning-based tool that, while not previously applied to optic nerve tissue, is widely used for nerve fiber segmentation.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>From 2,036 records, four manuscripts describing three deep learning models met inclusion criteria. Published correlation coefficients between model predictions and reference counts ranged from 0.959 to 0.99. On independent validation, performance was reduced: AxoNet 2.0 achieved the highest correlation (r = 0.89), followed by AxonDeepSeg (r = 0.86) and AxoNet (r = 0.79). Segmentation quality metrics revealed high precision (&gt;0.94) but low recall (0.18 to 0.27), with Dice coefficients of 0.29 to 0.40, substantially below published benchmarks of 0.81.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>Deep learning models for optic nerve histology demonstrate strong within-study performance but show meaningful performance decrements when applied to independent datasets. The observed generalizability gap (correlations 0.07 to 0.182 points below published values) demonstrates the need for standardized validation datasets and multi-center testing before widespread adoption of these tools.<\/jats:p>\n                <\/jats:sec>","DOI":"10.64898\/2026.03.11.710915","type":"posted-content","created":{"date-parts":[[2026,3,14]],"date-time":"2026-03-14T01:30:17Z","timestamp":1773451817000},"source":"Crossref","is-referenced-by-count":0,"title":["Comparison of Deep Learning Tools for Optic Nerve Axon Quantification Finds Limited Generalizability on Independent Validation"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-0354-0569","authenticated-orcid":false,"given":"Benton","family":"Chuter","sequence":"first","affiliation":[{"name":"Department of Ophthalmology, Hamilton Eye Institute, University of Tennessee Health Science Center, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-9370-9638","authenticated-orcid":false,"given":"Noah","family":"Emmert","sequence":"additional","affiliation":[{"name":"Miami University, Collierville, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-6747-0961","authenticated-orcid":false,"given":"Min Young","family":"Kim","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, Hamilton Eye Institute, University of Tennessee Health Science Center, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4232-4815","authenticated-orcid":false,"given":"Nikhil","family":"Dave","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, Hamilton Eye Institute, University of Tennessee Health Science Center, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-4492-8938","authenticated-orcid":false,"given":"Jay","family":"Herrin","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, Hamilton Eye Institute, University of Tennessee Health Science Center, Memphis, TN, USA"},{"name":"University of South Alabama Frederick P. Whiddon College of Medicine, Mobile, AL, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zirui","family":"Zhou","sequence":"additional","affiliation":[{"name":"University of Tennessee Health Science Center College of Medicine, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-8109-8350","authenticated-orcid":false,"given":"Gideon","family":"Wall","sequence":"additional","affiliation":[{"name":"University of Tennessee Health Science Center College of Medicine, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3634-0747","authenticated-orcid":false,"given":"Abraham","family":"Palmer","sequence":"additional","affiliation":[{"name":"Department of Psychiatry, University of California San Diego, San Diego, CA, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2680-6921","authenticated-orcid":false,"given":"Hao","family":"Chen","sequence":"additional","affiliation":[{"name":"Department of Pharmacology, Addiction Science, and Toxicology, The University of Tennessee Health Science Center, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2119-521X","authenticated-orcid":false,"given":"T.J.","family":"Hollingsworth","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, Hamilton Eye Institute, University of Tennessee Health Science Center, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4021-3559","authenticated-orcid":false,"given":"Monica M.","family":"Jablonski","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, Hamilton Eye Institute, University of Tennessee Health Science Center, Memphis, TN, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026031613151086000_2026.03.11.710915v1.1","doi-asserted-by":"publisher","DOI":"10.1016\/j.preteyeres.2011.11.002"},{"issue":"14","key":"2026031613151086000_2026.03.11.710915v1.2","doi-asserted-by":"crossref","first-page":"22","DOI":"10.1167\/tvst.10.14.22","article-title":"AxonDeep: Automated Optic Nerve Axon Segmentation in Mice with Deep Learning","volume":"10","year":"2021","journal-title":"Transl Vis Sci Technol"},{"issue":"4","key":"2026031613151086000_2026.03.11.710915v1.3","doi-asserted-by":"crossref","first-page":"686","DOI":"10.1038\/s41416-020-01122-x","article-title":"Deep learning in cancer pathology: a new generation of clinical biomarkers","volume":"124","year":"2021","journal-title":"Br J Cancer"},{"issue":"3","key":"2026031613151086000_2026.03.11.710915v1.4","doi-asserted-by":"crossref","first-page":"8","DOI":"10.1167\/tvst.12.3.8","article-title":"Morphometric Analysis of Retinal Ganglion Cell Axons in Normal and Glaucomatous Brown Norway Rats Optic Nerves","volume":"12","year":"2023","journal-title":"Transl Vis Sci Technol"},{"issue":"3","key":"2026031613151086000_2026.03.11.710915v1.5","doi-asserted-by":"crossref","first-page":"9","DOI":"10.1167\/tvst.12.3.9","article-title":"AxoNet 2.0: A Deep Learning-Based Tool for Morphometric Analysis of Retinal Ganglion Cell Axons","volume":"12","year":"2023","journal-title":"Transl Vis Sci Technol"},{"key":"2026031613151086000_2026.03.11.710915v1.6","doi-asserted-by":"publisher","DOI":"10.1109\/TBME.2021.3117407"},{"key":"2026031613151086000_2026.03.11.710915v1.7","doi-asserted-by":"publisher","DOI":"10.1038\/srep26286"},{"key":"2026031613151086000_2026.03.11.710915v1.8","first-page":"376","article-title":"Quantifying optic nerve axons in a cat glaucoma model by a semi-automated targeted counting method","volume":"20","year":"2014","journal-title":"Mol Vis"},{"key":"2026031613151086000_2026.03.11.710915v1.9","doi-asserted-by":"publisher","DOI":"10.1016\/j.exer.2015.06.006"},{"key":"2026031613151086000_2026.03.11.710915v1.10","doi-asserted-by":"publisher","DOI":"10.1167\/iovs.11-9274"},{"issue":"1","key":"2026031613151086000_2026.03.11.710915v1.11","doi-asserted-by":"crossref","first-page":"8034","DOI":"10.1038\/s41598-020-64898-1","article-title":"AxoNet: A deep learning-based tool to count retinal ganglion cell axons","volume":"10","year":"2020","journal-title":"Sci Rep"},{"key":"2026031613151086000_2026.03.11.710915v1.12","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2020.101813"},{"key":"2026031613151086000_2026.03.11.710915v1.13","doi-asserted-by":"publisher","DOI":"10.7326\/M18-0850"},{"issue":"9","key":"2026031613151086000_2026.03.11.710915v1.14","doi-asserted-by":"crossref","first-page":"1686","DOI":"10.1016\/j.ajpath.2019.05.007","article-title":"Pathology Image Analysis Using Segmentation Deep Learning Algorithms","volume":"189","year":"2019","journal-title":"Am J Pathol"},{"issue":"2","key":"2026031613151086000_2026.03.11.710915v1.15","doi-asserted-by":"crossref","first-page":"209","DOI":"10.1038\/eye.2016.295","article-title":"Assessing retinal ganglion cell damage","volume":"31","year":"2017","journal-title":"Eye (Lond)"},{"key":"2026031613151086000_2026.03.11.710915v1.16","doi-asserted-by":"publisher","DOI":"10.3389\/fgene.2020.00031"},{"key":"2026031613151086000_2026.03.11.710915v1.17","doi-asserted-by":"publisher","DOI":"10.1038\/srep26559"},{"issue":"1","key":"2026031613151086000_2026.03.11.710915v1.18","doi-asserted-by":"crossref","first-page":"3816","DOI":"10.1038\/s41598-018-22181-4","article-title":"AxonDeepSeg: automatic axon and myelin segmentation from microscopy data using convolutional neural networks","volume":"8","year":"2018","journal-title":"Sci Rep"},{"key":"2026031613151086000_2026.03.11.710915v1.19","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pmed.1002683"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.03.11.710915","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,16]],"date-time":"2026-03-16T20:15:19Z","timestamp":1773692119000},"score":1,"resource":{"primary":{"URL":"http:\/\/biorxiv.org\/lookup\/doi\/10.64898\/2026.03.11.710915"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,13]]},"references-count":19,"URL":"https:\/\/doi.org\/10.64898\/2026.03.11.710915","relation":{},"subject":[],"published":{"date-parts":[[2026,3,13]]},"subtype":"preprint"}}