Multi-input convolutional neural network for breast cancer detection using thermal images and clinical data

Sánchez-Cauce, Raquel; Pérez-Martín, Jorge; Luque, Manuel

doi:10.1016/j.cmpb.2021.106045

Cited by 74 publications

(40 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The severity of acute and late skin side effects of breast cancer irradiation is dependent on many physical and clinical factors, especially the total radiation dose, fractionation, irradiated volume, patient’s age, comorbidities and genetic predisposition. In the course of radiotherapy, skin toxicity remains an important clinical problem for many patients [ 41 , 42 , 43 , 44 , 45 , 46 , 47 ].…”

Section: Discussionmentioning

confidence: 99%

Long-Term Skin Temperature Changes after Breast Cancer Radiotherapy

Baic

Plaza

Lange

et al. 2022

IJERPH

View full text Add to dashboard Cite

The aim of the study was to use thermal imaging to evaluate long-term chest temperature changes in patients who had previously been treated with radiotherapy. The examination with a thermal imaging camera involved 144 women—48 of them were patients after RT, 48 were females before breast cancer radiotherapy and the last group of participants were 48 healthy women. All patients (before and after radiotherapy) were divided into women after mastectomy and those after conservative surgery. In addition, the first group of women, those who had received radiotherapy, were divided into three other groups: up to 1 year after RT, over 1 year and up to 5 years after RT and over 5 years after RT. Due to this, it was possible to compare the results and analyse the differences between the temperature in the healthy and treated breasts. The comparison of obtained temperature results showed that the area treated by ionizing radiation is characterized by a higher temperature even a few years after the finished treatment. It is worth mentioning that despite the fact that the difference was visible on the thermograms, the patients had no observable skin lesion or change in color at the treatment site. For the results of the study provided for the group of healthy patients, there were no significant differences observed between the average temperatures in the breasts. The use of thermal imaging in the evaluation of skin temperature changes after radiotherapy showed that the average temperature in the treated breast area can change even a long time after treatment.

show abstract

Section: Discussionmentioning

confidence: 99%

Long-Term Skin Temperature Changes after Breast Cancer Radiotherapy

Baic

Plaza

Lange

et al. 2022

IJERPH

View full text Add to dashboard Cite

show abstract

“…The performance of the model was improved using data augmentation techniques and global contrast normalization. In addition, in [ 42 ], the authors proposed a classification method for thermal images, which combines thermal images of different views using a CNN model. The method achieved 97% accuracy and 0.99 AUC, with a 100% specificity and 83% sensitivity.…”

Section: Literature Reviewmentioning

confidence: 99%

An optimized deep learning architecture for breast cancer diagnosis based on improved marine predators algorithm

Houssein

Emam

Ali

2022

Neural Comput & Applic

View full text Add to dashboard Cite

Breast cancer is the second leading cause of death in women; therefore, effective early detection of this cancer can reduce its mortality rate. Breast cancer detection and classification in the early phases of development may allow for optimal therapy. Convolutional neural networks (CNNs) have enhanced tumor detection and classification efficiency in medical imaging compared to traditional approaches. This paper proposes a novel classification model for breast cancer diagnosis based on a hybridized CNN and an improved optimization algorithm, along with transfer learning, to help radiologists detect abnormalities efficiently. The marine predators algorithm (MPA) is the optimization algorithm we used, and we improve it using the opposition-based learning strategy to cope with the implied weaknesses of the original MPA. The improved marine predators algorithm (IMPA) is used to find the best values for the hyperparameters of the CNN architecture. The proposed method uses a pretrained CNN model called ResNet50 (residual network). This model is hybridized with the IMPA algorithm, resulting in an architecture called IMPA-ResNet50. Our evaluation is performed on two mammographic datasets, the mammographic image analysis society (MIAS) and curated breast imaging subset of DDSM (CBIS-DDSM) datasets. The proposed model was compared with other state-of-the-art approaches. The obtained results showed that the proposed model outperforms the compared state-of-the-art approaches, which are beneficial to classification performance, achieving 98.32% accuracy, 98.56% sensitivity, and 98.68% specificity on the CBIS-DDSM dataset and 98.88% accuracy, 97.61% sensitivity, and 98.40% specificity on the MIAS dataset. To evaluate the performance of IMPA in finding the optimal values for the hyperparameters of ResNet50 architecture, it compared to four other optimization algorithms including gravitational search algorithm (GSA), Harris hawks optimization (HHO), whale optimization algorithm (WOA), and the original MPA algorithm. The counterparts algorithms are also hybrid with the ResNet50 architecture produce models named GSA-ResNet50, HHO-ResNet50, WOA-ResNet50, and MPA-ResNet50, respectively. The results indicated that the proposed IMPA-ResNet50 is achieved a better performance than other counterparts.

show abstract

“…Further studies have been done in the localization of such regions which helps radiologists study it further and provide the necessary treatments. In [ 6 ], the authors have tried machine learning algorithms to predict the presence of cancer using thermal images of the breast. The most common screening technique currently used is mammography which has disadvantages such as exposure to radiation, high costs to get a screening, and discomfort to the patient.…”

Section: Problem Statementmentioning

confidence: 99%

“…The most common screening technique currently used is mammography which has disadvantages such as exposure to radiation, high costs to get a screening, and discomfort to the patient. Paper [ 6 ] is among the first few papers that worked on adding clinical data of the patient to the multi-input convolutional neural network.…”

Section: Problem Statementmentioning

confidence: 99%

“…The reason for picking thermography over mammography is, as mentioned in the report, the boom of AI opened new possibilities, and thermography being a contact-less screening method makes it more preferred over mammography. The idea behind adding clinical data along with the thermal images is that it enhances the accuracy of the model, as exhibited in [ 6 ]. Table 3 shows some pros and cons regarding certain other approaches discussed earlier.…”

Section: Literature Reviewmentioning

confidence: 99%

See 1 more Smart Citation

Detection of Breast Cancer from Five-View Thermal Images Using Convolutional Neural Networks

Mammoottil

Kulangara

Cherian

et al. 2022

Journal of Healthcare Engineering

View full text Add to dashboard Cite

Breast cancer is one of the most common forms of cancer. Its aggressive nature coupled with high mortality rates makes this cancer life-threatening; hence early detection gives the patient a greater chance of survival. Currently, the preferred diagnosis method is mammography. However, mammography is expensive and exposes the patient to radiation. A cost-effective and less invasive method known as thermography is gaining popularity. Bearing this in mind, the work aims to initially create machine learning models based on convolutional neural networks using multiple thermal views of the breast to detect breast cancer using the Visual DMR dataset. The performances of these models are then verified with the clinical data. Findings indicate that the addition of clinical data decisions to the model helped increase its performance. After building and testing two models with different architectures, the model used the same architecture for all three views performed best. It performed with an accuracy of 85.4%, which increased to 93.8% after the clinical data decision was added. After the addition of clinical data decisions, the model was able to classify more patients correctly with a specificity of 96.7% and sensitivity of 88.9% when considering sick patients as the positive class. Currently, thermography is among the lesser-known diagnosis methods with only one public dataset. We hope our work will divert more attention to this area.

show abstract

Multi-input convolutional neural network for breast cancer detection using thermal images and clinical data

Cited by 74 publications

References 29 publications

Long-Term Skin Temperature Changes after Breast Cancer Radiotherapy

Long-Term Skin Temperature Changes after Breast Cancer Radiotherapy

An optimized deep learning architecture for breast cancer diagnosis based on improved marine predators algorithm

Detection of Breast Cancer from Five-View Thermal Images Using Convolutional Neural Networks

Contact Info

Product

Resources

About