| Title: |
Spatial-Spectral Deep Learning for Prostate Cancer Tissue Classification in Infrared Spectroscopy |
| Authors: |
O'Leary, Lyra; Ferguson, Dougal; Hart, Claire Alexandra; Brown, Michael; Oliveira, Pedro; Clarke, Noel; Sachdeva, Ashwin; Gardner, Peter; Yin, Hujun |
| Source: |
O'Leary, L, Ferguson, D, Hart, C A, Brown, M, Oliveira, P, Clarke, N, Sachdeva, A, Gardner, P & Yin, H 2026, 'Spatial-Spectral Deep Learning for Prostate Cancer Tissue Classification in Infrared Spectroscopy', Analytical Chemistry, vol. 98, no. 4, PMID 370536, pp. 2743–2755. https://doi.org/10.1021/acs.analchem.5c04765 |
| Publication Year: |
2026 |
| Collection: |
The University of Manchester: Research Explorer - Publications |
| Description: |
Modern methods of infrared (IR) spectroscopy yield full IR absorbance spectra in arrays, forming hyperspectral images. End-to-end processing of these images via deep learning seems ideal for exploiting their high dimensionality and wealth of spatial and spectral information, but recent research suggests that convolution-based architectures may have a spatial bias. Towards the goal of improved prostate cancer tissue classification, we compare a variety of deep learning classifiers for IR spectroscopy and probe the impact of a bottleneck which compresses the spectral dimension. We find a strong correlation between model spatial receptive field and classification performance, with highest performance achieved by a modified Vision Transformer model. Conversely, we find only limited correlation between spectral information and deep learning model performance: we find that a spectral bottleneck of just 16 features has only a negligible effect on all neural network models, including convolution-eschewing transformer architectures and a multi-layer perceptron model utilising no spatial information. Rather than any particular network component inducing a spatial bias, the breadth of architectures exhibiting little dependence on spectral information implies that tissue classification itself is characterised by only a small set of spectral features. This in turn suggests that success at tissue classification may be a poor benchmark in the development of deep learning models designed to effectively utilise the spectral dimension. |
| Document Type: |
article in journal/newspaper |
| File Description: |
application/pdf |
| Language: |
English |
| ISSN: |
0003-2700; 1520-6882 |
| Relation: |
info:eu-repo/semantics/altIdentifier/pissn/0003-2700; info:eu-repo/semantics/altIdentifier/eissn/1520-6882 |
| DOI: |
10.1021/acs.analchem.5c04765 |
| Availability: |
https://research.manchester.ac.uk/en/publications/b38d4e6d-76e9-453d-a2f3-e1c95583b965; https://doi.org/10.1021/acs.analchem.5c04765; https://pure.manchester.ac.uk/ws/files/1773921257/final_manuscript.pdf |
| Rights: |
info:eu-repo/semantics/openAccess ; http://creativecommons.org/licenses/by/4.0/ |
| Accession Number: |
edsbas.EACAF250 |
| Database: |
BASE |