Assessing Rectal Cancer Treatment Response Using Coregistered Endorectal Photoacoustic and US Imaging Paired with Deep Learning﻿

Leng, Xiandong; Uddin, K. M. Shihab; Chapman, William C.; Luo, Hongbo; Kou, Sitai; Amidi, Eghbal; Yang, Guang; Chatterjee, Deyali; Shetty, Anup S.; Hunt, S.; Mutch, Matthew G.; Zhu, Quing

doi:10.1148/radiol.2021202208

Cited by 20 publications

(21 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have observed uniform, layer-like vasculature with intense photoacoustic signals within normal rectal submucosa and in the tumor beds where complete tumor destruction has occurred. In contrast, heterogeneous and often microvascular-deficient regions have been found consistently in tumor beds with residual cancer at treatment completion ( 13 , 14 ). The return of a “normal” vascular pattern to the tumor bed appears to signal complete tumor destruction, though this mechanism is not well-understood.…”

Section: Discussionmentioning

confidence: 99%

“…The PAM endoscope consists three parts: a handle, a water channel (the main body), and an imaging head, as shown in Figure 1A ( 14 ). Briefly, the water inlet which allows water injected from a syringe to inflate a water balloon covering the image head for ultrasound coupling.…”

Section: Methodsmentioning

confidence: 99%

“…The PAM-CNN (or US-CNN) architecture ( Figure 5 ) contained two sequential feature extraction layers and two fully connected layers ( 14 ). Briefly, each extraction layer had a convolutional layer followed by a pooling layer.…”

Section: Methodsmentioning

confidence: 99%

“…To overcome these challenges, we developed a co-registered endorectal photoacoustic microscopy and ultrasound (PAM/US) system to assess rectal cancer treatment response ( 13 , 14 ). Photoacoustic imaging (PAI) is a hybrid imaging technology that uses a short laser pulses to excite hemoglobin molecules endogenous to the human body.…”

Section: Introductionmentioning

confidence: 99%

“…A convolutional neural network (CNN) is an artificial intelligence algorithm with remarkable capabilities for automated image analysis ( 23 ). To quantitatively interpret the large volumes of data acquired by the PAM/US system, we designed and incorporated deep-learning CNN models in the PAM system (PAM-CNN) ( 14 ). While our deep-learning PAM-CNN model can accurately assess rectal cancer treatment response, it requires a large training and validation data set.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Rectal Cancer Treatment Management: Deep-Learning Neural Network Based on Photoacoustic Microscopy Image Outperforms Histogram-Feature-Based Classification

et al. 2021

Self Cite

View full text Add to dashboard Cite

We have developed a novel photoacoustic microscopy/ultrasound (PAM/US) endoscope to image post-treatment rectal cancer for surgical management of residual tumor after radiation and chemotherapy. Paired with a deep-learning convolutional neural network (CNN), the PAM images accurately differentiated pathological complete responders (pCR) from incomplete responders. However, the role of CNNs compared with traditional histogram-feature based classifiers needs further exploration. In this work, we compare the performance of the CNN models to generalized linear models (GLM) across 24 ex vivo specimens and 10 in vivo patient examinations. First order statistical features were extracted from histograms of PAM and US images to train, validate and test GLM models, while PAM and US images were directly used to train, validate, and test CNN models. The PAM-CNN model performed superiorly with an AUC of 0.96 (95% CI: 0.95-0.98) compared to the best PAM-GLM model using kurtosis with an AUC of 0.82 (95% CI: 0.82-0.83). We also found that both CNN and GLMs derived from photoacoustic data outperformed those utilizing ultrasound alone. We conclude that deep-learning neural networks paired with photoacoustic images is the optimal analysis framework for determining presence of residual cancer in the treated human rectum.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Rectal Cancer Treatment Management: Deep-Learning Neural Network Based on Photoacoustic Microscopy Image Outperforms Histogram-Feature-Based Classification

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

Characterization of adnexal lesions using photoacoustic imaging to improve sonographic O‐RADS risk assessment

Zhu,

Luo,

Middleton

et al. 2023

Ultrasound in Obstet & Gyne

Self Cite

View full text Add to dashboard Cite

ObjectiveTo assess the impact of photoacoustic imaging (PAI) on improving ultrasound (US) O‐RADS assessment in women undergoing oophorectomy who have an adnexal lesion(s) suggestive of malignancy.MethodThis prospective study enrolled women referred for oophorectomy. Participants underwent clinical US followed by co‐registered US and photoacoustic imaging (PAI) before oophorectomy. Each ovarian/adnexal lesion was graded by two radiologists using the US O‐RADS scale. PAI was used to compute relative total hemoglobin concentration (rHbT) and blood oxygenation saturation (%sO2) maps in the region of interest. Lesions were categorized by histopathology into malignant ovarian/ adnexal lesion, malignant fallopian tube, and several benign categories to assess the impact of PAI in assisting O‐RADS for risk assessment. rHbT and %sO2 were compared between malignant and benign histologic groups using a one‐way ANOVA test followed by a post‐hoc Dunnett test of significance. Logistic‐regression models were developed based on tumor marker CA125 alone, US‐based O‐RADS alone, PAI rHbT and %sO2. AUC under ROC was used to compare the diagnostic performance of the models.ResultsThere were 93 lesions among 68 women (mean age 52 years, range 21–79). Surgical pathology revealed 14 patients with malignant ovarian/ adnexal lesions, 2 patients with malignant fallopian tubes, and 52 patients with benign findings. rHbT was significantly different between malignant lesions and benign cases; %sO2 was lower in malignant lesions, but not statistically significant in all benign categories. Feature analysis revealed that rHbT, O‐RADS, CA125, and %sO2 were the most important predictors of malignancy. Logistic‐regression models revealed AUC=0.79 (95% CI: 0.626‐0.953) for CA125 alone, AUC=0.86 (95% CI: 0.732‐0.981) for O‐RADS only, AUC=0.88 (95% CI: 0.7596‐1) for CA125 and O‐RADS, AUC=0.90 (95% CI: 0.815‐0.985) for rHbT and %sO2. A model utilizing all four features (CA125, O‐RADS, rHbT, and %sO2) achieved a superior performance of AUC=0.97 (95% CI: 0.932‐1) with sensitivity of 100% and specificity of 82%.ConclusionsThe additional PAI rHbT and %sO2 information significantly improved the US O‐RADS in the accurate diagnosis of adnexal lesions.This article is protected by copyright. All rights reserved.

show abstract

Cancer detection and classification using a simplified binary state vector machine

Shafi,

Ansari,

Din

et al. 2024

Med Biol Eng Comput

View full text Add to dashboard Cite

Assessing Rectal Cancer Treatment Response Using Coregistered Endorectal Photoacoustic and US Imaging Paired with Deep Learning

Cited by 20 publications

References 23 publications

Rectal Cancer Treatment Management: Deep-Learning Neural Network Based on Photoacoustic Microscopy Image Outperforms Histogram-Feature-Based Classification

Rectal Cancer Treatment Management: Deep-Learning Neural Network Based on Photoacoustic Microscopy Image Outperforms Histogram-Feature-Based Classification

Characterization of adnexal lesions using photoacoustic imaging to improve sonographic O‐RADS risk assessment

Cancer detection and classification using a simplified binary state vector machine

Contact Info

Product

Resources

About

Assessing Rectal Cancer Treatment Response Using Coregistered Endorectal Photoacoustic and US Imaging Paired with Deep Learning﻿

Cited by 20 publications

References 23 publications

Rectal Cancer Treatment Management: Deep-Learning Neural Network Based on Photoacoustic Microscopy Image Outperforms Histogram-Feature-Based Classification

Rectal Cancer Treatment Management: Deep-Learning Neural Network Based on Photoacoustic Microscopy Image Outperforms Histogram-Feature-Based Classification

Characterization of adnexal lesions using photoacoustic imaging to improve sonographic O‐RADS risk assessment

Cancer detection and classification using a simplified binary state vector machine

Contact Info

Product

Resources

About

Assessing Rectal Cancer Treatment Response Using Coregistered Endorectal Photoacoustic and US Imaging Paired with Deep Learning