Morphological pattern classification system for plantar thermography of patients with diabetes

Update on the Use of Infrared Thermography in the Early Detection of Diabetic Foot Complications: A Bibliographic Review

Foot lesions are among the most frequent causes of morbidity and disability in the diabetic population. Thus, the exploration of preventive control measures is vital for detecting early signs and symptoms of this disease. Infrared thermography is one of the complementary diagnostic tools available that has proven to be effective in the control of diabetic foot. The last review on this topic was published in 2015 and so, we conducted a bibliographic review of the main databases (PubMed, the Web of Science, Cochrane library, and Scopus) during the third quarter of 2023. We aimed to identify the effectiveness of infrared thermography as a diagnostic element in pre-ulcerous states in diabetic patients and to detect diabetic foot ulcer complications. We obtained a total of 1199 articles, 26 of which were finally included in the present review and published after 2013. After analyzing the use of infrared thermography in diabetic patients both with and without ulcers, as well as in healthy individuals, we concluded that is an effective tool for detecting early-stage ulcers in diabetic foot patients.

Region-wise severity analysis of diabetic plantar foot thermograms

OBJECTIVES

Diabetic foot ulcers (DFU) can be avoided if symptoms of diabetic foot complications are detected early and treated promptly. Early detection requires regular examination, which might be limited for many reasons. To identify affected or potentially affected regions in the diabetic plantar foot, the region-wise severity of the plantar foot must be known.

METHODS

A novel thermal diabetic foot dataset of 104 subjects was developed that is suitable for Indian healthcare conditions. The entire plantar foot thermogram is divided into three parts, i.e., forefoot, midfoot, and hindfoot. The division of plantar foot is based on the prevalence of foot ulcers and the load on the foot. To classify the severity levels, conventional machine learning (CML) techniques like logistic regression, decision tree, KNN, SVM, random forest, etc., and convolutional neural networks (CNN), such as EfficientNetB1, VGG-16, VGG-19, AlexNet, InceptionV3, etc., were applied and compared for robust outcomes.

RESULTS

The study successfully developed a thermal diabetic foot dataset, allowing for effective classification of diabetic foot ulcer severity using the CML and CNN techniques. The comparison of different methods revealed variations in performance, with certain approaches outperforming others.

CONCLUSIONS

The region-based severity analysis offers valuable insights for targeted interventions and preventive measures, contributing to a comprehensive assessment of diabetic foot ulcer severity. Further research and development in these techniques can enhance the detection and management of diabetic foot complications, ultimately improving patient outcomes.

The effectiveness of a thermography-driven preventive foot care protocol on the recurrence of diabetic foot ulcers in low-medical resource settings: An open-labeled randomized controlled trial

BACKGROUND

Plantar temperature monitoring, along with a comprehensive preventive foot care approach, is utilized to prevent recurrence of diabetic foot ulcers. However, there is a lack of standardized protocols for individuals with diabetic foot ulcer history in low-medical resource countries.

OBJECTIVE

This study investigated the efficacy of nurse-led, thermographic-evaluation-guided foot care in preventing diabetic foot ulcer recurrence in a low-medical resource country.

DESIGN

Single-blind, 1:1 allocation randomized controlled trial.

SETTINGS

Two wound care facilities with wound care nurse specialists in Indonesia.

PARTICIPANTS

120 patients with a diabetic foot ulcer history.

INTERVENTION

In the intervention group, baseline risk assessment including smartphone thermography evaluation was performed. Personalized foot care and education were conducted monthly for participants whose thermographs showed increased foot lesion temperature at baseline risk assessment. The control group received usual care and education using a booklet at baseline. The follow-up period was six months.

MAIN OUTCOME

Time to recurrence was evaluated using Kaplan-Meier survival analysis, and between-group comparisons were performed using the log-rank test. Potential risk factors were incorporated into the multivariate Cox regression model. Secondary outcomes included quality of life (European Quality of Life 5 Dimensions 3 Level Version) and foot care behavior, were analyzed using Mixed Models for Repeated Measures at baseline, third, and sixth follow-up (3 and 6 months from baseline).

RESULTS

120 participants (intervention 60, control 60) were randomized. The intervention group had a significantly lower recurrence rate than the control group (15% vs. 35%, p = 0.011) and a significant difference in time to ulceration (log-rank test, p = 0.009) after a 6-month follow-up period. Intervention care reduced the risk of diabetic foot ulcer recurrence by 59% (Hazard ratio 0.41, 95% confidence interval 0.18, 0.96, p = 0.039) in multivariate Cox regression analysis. On the third follow-up, total diabetic foot care behavior score (p < 0.001) was significantly improved by the intervention. On the sixth follow-up, mobility (p = 0.020), self-care (p = 0.023), pain/discomfort (p < 0.001), anxiety/depression (p = 0.016), EuroQol Visual Analogue Scale score (p = 0.002), and total diabetic foot care behavior score (p < 0.001) showed significant improvements in the intervention group.

CONCLUSIONS

Foot care and personalized education delivered at a frequency based on the risk level assessed by thermography effectively reduced diabetic foot ulcer recurrence, and improved quality of life and foot care behaviors.

REGISTRATION NUMBER

UMIN000039012.

TWEETABLE ABSTRACT

Nurse-led diabetic foot care and education reduced diabetic foot ulcer recurrence in Indonesia.

Assessment of chronic limb threatening ischemia using thermal imaging

OBJECTIVES

Current chronic limb threatening ischemia (CLTI) diagnostics require expensive equipment, using ionizing radiation or contrast agents, or summative surrogate methods lacking in spatial information. Our aim is to develop and improve contactless, non-ionizing and cost-effective diagnostic methods for CLTI assessment with high spatial accuracy by utilizing dynamic thermal imaging and the angiosome concept.

APPROACH

Dynamic thermal imaging test protocol was suggested and implemented with a number of computational parameters. Pilot data was measured from 3 healthy young subjects, 4 peripheral artery disease (PAD) patients and 4 CLTI patients. The protocol consists of clinical reference measurements, including ankle- and toe-brachial indices (ABI, TBI), and a modified patient bed for hydrostatic and thermal modulation tests. The data was analyzed using bivariate correlation.

RESULTS

The thermal recovery time constant was on average higher for the PAD (88%) and CLTI (83%) groups with respect to the healthy young subjects. The contralateral symmetry was high for the healthy young group and low for the CLTI group. The recovery time constants showed high negative correlation to TBI (ρ = -0.73) and ABI (ρ = -0.60). The relation of these clinical parameters to the hydrostatic response and absolute temperatures (|ρ|<0.3) remained unclear.

CONCLUSION

The lack of correlation for absolute temperatures or their contralateral differences with the clinical status, ABI and TBI disputes their use in CLTI diagnostics. Thermal modulation tests tend to augment the signs of thermoregulation deficiencies and accordingly high correlations were found with all reference metrics. The method is promising for establishing the connection between impaired perfusion and thermography. The hydrostatic modulation test requires more research with stricter test conditions.

A Novel Machine Learning Approach for Severity Classification of Diabetic Foot Complications Using Thermogram Images

Diabetes mellitus (DM) is one of the most prevalent diseases in the world, and is correlated to a high index of mortality. One of its major complications is diabetic foot, leading to plantar ulcers, amputation, and death. Several studies report that a thermogram helps to detect changes in the plantar temperature of the foot, which may lead to a higher risk of ulceration. However, in diabetic patients, the distribution of plantar temperature does not follow a standard pattern, thereby making it difficult to quantify the changes. The abnormal temperature distribution in infrared (IR) foot thermogram images can be used for the early detection of diabetic foot before ulceration to avoid complications. There is no machine learning-based technique reported in the literature to classify these thermograms based on the severity of diabetic foot complications. This paper uses an available labeled diabetic thermogram dataset and uses the k-mean clustering technique to cluster the severity risk of diabetic foot ulcers using an unsupervised approach. Using the plantar foot temperature, the new clustered dataset is verified by expert medical doctors in terms of risk for the development of foot ulcers. The newly labeled dataset is then investigated in terms of robustness to be classified by any machine learning network. Classical machine learning algorithms with feature engineering and a convolutional neural network (CNN) with image-enhancement techniques are investigated to provide the best-performing network in classifying thermograms based on severity. It is found that the popular VGG 19 CNN model shows an accuracy, precision, sensitivity, F1-score, and specificity of 95.08%, 95.08%, 95.09%, 95.08%, and 97.2%, respectively, in the stratification of severity. A stacking classifier is proposed using extracted features of the thermogram, which is created using the trained gradient boost classifier, XGBoost classifier, and random forest classifier. This provides a comparable performance of 94.47%, 94.45%, 94.47%, 94.43%, and 93.25% for accuracy, precision, sensitivity, F1-score, and specificity, respectively.

Prevention of diabetic foot ulcers using a smartphone and mobile thermography: a case study

OBJECTIVE

Early identification of pre-ulcerative pathology is important to preventing diabetic foot ulcers (DFU), but signs of inflammation are difficult to detect on the feet of patients with diabetic neuropathy due to decreased sensation. However, infrared thermography can objectively identify inflammation. Therefore, a device that allows patients to visualise thermograms of their feet might be an effective way to prevent DFU. We aimed to determine the effects of a novel self-monitoring device to prevent DFU using a thermograph attached to a smartphone.

METHOD

A self-monitoring device comprising a mobile thermograph attached to a smartphone on a selfie stick was created, and its effects in two patients with diabetic neuropathy and foot calluses assessed.

RESULTS

For one patient, he understood that walking too much increased the temperature in the skin of his feet (a sign of inflammation). The other patient could not detect high-risk findings, because the temperature of his skin did not increase during the study period.

CONCLUSION

This device might provide self-care incentives to prevent DFU, although some issues, such as the automatic detection of high-risk thermographic changes, need to be improved.

Effectiveness of a Central Discharge Element Sock for Plantar Temperature Reduction and Improving Comfort

U-shaped plantar cushions could help reduce stress affecting the central forefoot without the need for an orthosis, but they are yet to be integrated as an element in socks. The objective of this study was to verify the effectiveness of a sock with a central discharge element in terms of plantar temperature and comfort. The sample comprised 38 subjects (13 men and 25 women). Their plantar temperatures were measured with a thermographic camera in a basal situation and after each of two 10-minute walks around an indoor circuit during which they wore either control or experimental socks at random (the same design, weight, and fiber, but with the plantar cushioning element added). After the walks, each subject responded to a comfort questionnaire (five-point Likert scale), blindly scoring the two socks. The highest temperatures (28.3 ± 2.7 °C) were recorded in the zone of the second and third metatarsal heads. With the experimental socks, the observed temperature increase in the central forefoot zone was significantly less than with the control socks (31.6 vs. 30.6 °C, p = 0.001). The subjects found the experimental socks to be more comfortable than the controls (4.63 ± 0.5 vs. 4.03 ± 0.5, p < 0.001). The discharge element included in the experimental socks was effective since it reduced the contact zones and excess friction with the ground, thereby lessening overheating by more than 1 °C. Furthermore, the experimental socks were perceived as being more comfortable by the subjects who had mild and occasional foot discomfort.

Deep Learning Classification for Diabetic Foot Thermograms

According to the World Health Organization (WHO), Diabetes Mellitus (DM) is one of the most prevalent diseases in the world. It is also associated with a high mortality index. Diabetic foot is one of its main complications, and it comprises the development of plantar ulcers that could result in an amputation. Several works report that thermography is useful to detect changes in the plantar temperature, which could give rise to a higher risk of ulceration. However, the plantar temperature distribution does not follow a particular pattern in diabetic patients, thereby making it difficult to measure the changes. Thus, there is an interest in improving the success of the analysis and classification methods that help to detect abnormal changes in the plantar temperature. All this leads to the use of computer-aided systems, such as those involved in artificial intelligence (AI), which operate with highly complex data structures. This paper compares machine learning-based techniques with Deep Learning (DL) structures. We tested common structures in the mode of transfer learning, including AlexNet and GoogleNet. Moreover, we designed a new DL-structure, which is trained from scratch and is able to reach higher values in terms of accuracy and other quality measures. The main goal of this work is to analyze the use of AI and DL for the classification of diabetic foot thermograms, highlighting their advantages and limitations. To the best of our knowledge, this is the first proposal of DL networks applied to the classification of diabetic foot thermograms. The experiments are conducted over thermograms of DM and control groups. After that, a multi-level classification is performed based on a previously reported thermal change index. The high accuracy obtained shows the usefulness of AI and DL as auxiliary tools to aid during the medical diagnosis.

Telehealth and telemedicine applications for the diabetic foot: A systematic review

The aim of this systematic review is to assess the peer-reviewed literature on the psychometric properties, feasibility, effectiveness, costs, and current limitations of using telehealth and telemedicine approaches for prevention and management of diabetic foot disease. MEDLINE/PubMed was searched for peer-reviewed studies on telehealth and telemedicine approaches for assessing, monitoring, preventing, or treating diabetic foot disease. Four modalities were formulated: dermal thermography, hyperspectral imaging, digital photographic imaging, and audio/video/online communication. Outcome measures were: validity, reliability, feasibility, effectiveness, and costs. Sixty-one studies were eligible for analysis. Three randomized controlled trials showed that handheld infrared dermal thermography as home-monitoring tool is effective in reducing ulcer recurrence risk, while one small trial showed no effect. Hyperspectral imaging has been tested in clinical settings to assess and monitor foot disease and conflicting results on its diagnostic use show that this method is still in an experimental stage. Digital photography is used to assess and monitor foot ulcers and pre-ulcerative lesions and was found to be a valid, reliable, and feasible method for telehealth purposes. Audio/video/online communication is mainly used for foot ulcer monitoring. Two randomized controlled trials show similar healing efficacy compared with regular outpatient clinic visits, but no benefit in costs. In conclusion, several technologies with good psychometric properties are available that may be of benefit in helping to assess, monitor, prevent, or treat diabetic foot disease, but in most cases, feasibility, effectiveness, and cost savings still need to be demonstrated to become accepted and used modalities in diabetic foot care.

Relationship between skin temperature and soft tissue hardness in diabetic patients: an exploratory study

OBJECTIVE

The role of skin temperature and soft tissue hardness in the development of plantar ulcers is still in debate. However, the relationship between skin temperature and soft tissue hardness has not been explored. This study intends to analyse an eventual association between skin temperature and soft tissue hardness in the foot of diabetic patients Approach: Twenty diabetic patients enrolled for this study. The analysis was done at the foot level, therefore, skin temperature and soft tissue hardness data of the plantar surface of 40 feet were obtained in eight regions of the foot, two in the heel, two in the midfoot, three in the forefoot and one in the hallux. Information regarding glycaemic control (HbA1c levels) was retrieved from the clinical records of the patients.

MAIN RESULTS

After averaging skin temperature and soft tissue hardness in the calcaneum (medial and lateral), in the midfoot (medial and lateral) and in the metatarsal head (1st, 2nd-3rd and 4th-5th), a negative, moderate and significant association was found between skin temperature and soft tissue hardness in the metatarsal head (rho  =  -0.553; p   <  0.001), a positive, low and significant association was found in the midfoot (rho  =  0.333; p   =  0.036), but no association was found in the heel. The multiple linear regression models with skin temperature as dependent variable and soft tissue hardness as predictor were statistically significant in the metatarsal heads and midfoot, and explained 28.8% (R ²  =  0.288, F _(1,38)  =  15.37, p   <  0.001) and 11.9% (R ²  =  0.119, F _(1,38)  =  5.151, p   =  0.029) of the variance in skin temperature, respectively.

SIGNIFICANCE

Skin temperature is negatively associated with soft tissue hardness in the metatarsal head region and positively associated with soft tissue hardness in the midfoot. These findings imply that soft tissue hardness should be considered in the assessment of diabetic foot patients and that this variable should be controlled in studies assessing the determinants of foot skin temperature.

Reliability of Infrared Thermography Images in the Analysis of the Plantar Surface Temperature in Diabetes Mellitus

Objective

The purpose of this study was to evaluate the intraexaminer and interexaminer reliability of image assessment of infrared thermography of the plantar surface of people with diabetes mellitus.

Methods

Fifty-one participants with diabetes were included. The interexaminer reliability (test) consisted of measuring the temperature of the plantar surface by 2 evaluators who separately performed the analysis of thermographic images. The intraexaminer reliability (retest) consisted of the revaluation of the images by a single evaluator 10 days after the first analysis.

Results

The analysis of intrareliability and interreliability indicated excellent levels of correlation (intraclass correlation coefficient >0.75, 95% confidence interval >0.70) in the plantar surface points analyzed.

Conclusion

Infrared thermography image evaluation identified intrareliability and interreliability for temperature analysis of the plantar surface of people with diabetes mellitus. However, further studies need to be conducted to assess validity and its application in health care.

The thermoregulation of healthy individuals, overweight-obese, and diabetic from the plantar skin thermogram: a clue to predict the diabetic foot

Background: Thermoregulation is a complex autonomic process to keep or to dissipate heat in the human body.

Methods: In this work, by means of the thermogram of the plantar skin, the thermoregulation of healthy individuals, overweight–obese, and diabetic is discussed.

Results: The thermograms of the plantar skin, for the healthy individuals, are: (1) symmetrical, the temperature distribution of the right foot being a mirror image of that of the left foot ; (2) the thermograms of women, on average, are 3°C colder than those of the men; and (3) the temperature distributions decrease distally from the medial longitudinal arch. The plantar skin thermograms of overweight–obese individuals show: (1) increased average temperature of both feet and for both genders; (2) no symmetry between the left and right feet thermograms; and (3) the temperature distribution is still decreasing from the medial longitudinal arch to the periphery of the foot. However, the standard deviation, for each averaged temperature of the angiosomes, shows greater uncertainty. Most thermograms of diabetic individuals show temperature increase on the plantar skin, and are mostly symmetric between left and right feet.

Conclusions: An asymmetric thermogram of the plantar skin of diabetic individuals, where one foot is hotter than the other, may mean that the coldest foot is losing the capacity to communicate properly with the central nervous system and/or that vasoconstriction/vasodilatation is having problems in regulating the passing of blood through the vessels. Thus, the asymmetric thermograms of diabetic patients, and particularly those coldest regions of foot are of interest, because of the reduction of the local autonomic sensing and the lack of achieving properly the passing of the blood.

An early diagnostic tool for diabetic peripheral neuropathy in rats

The skin’s rewarming rate of diabetic patients is used as a diagnostic tool for early diagnosis of diabetic neuropathy. At present, the relationship between microvascular changes in the skin and diabetic neuropathy is unclear in streptozotocin (STZ) diabetic rats. The aim of this study was to investigate whether the skin rewarming rate in diabetic rats is related to microvascular changes and whether this is accompanied by changes observed in classical diagnostic methods for diabetic peripheral neuropathy. Computer-assisted infrared thermography was used to assess the rewarming rate after cold exposure on the plantar skin of STZ diabetic rats’ hind paws. Peripheral neuropathy was determined by the density of intra-epidermal nerve fibers (IENFs), mechanical sensitivity, and electrophysiological recordings. Data were obtained in diabetic rats at four, six, and eight weeks after the induction of diabetes and in controls. Four weeks after the induction of diabetes, a delayed rewarming rate, decreased skin blood flow and decreased density of IENFs were observed. However, the mechanical hyposensitivity and decreased motor nerve conduction velocity (MNCV) developed 6 and 8 weeks after the induction of diabetes. Our study shows that the skin rewarming rate is related to microvascular changes in diabetic rats. Moreover, the skin rewarming rate is a non-invasive method that provides more information for an earlier diagnosis of peripheral neuropathy than the classical monofilament test and MNCV in STZ induced diabetic rats.

Application of infrared thermography in computer aided diagnosis

The invention of thermography, in the 1950s, posed a formidable problem to the research community: What is the relationship between disease and heat radiation captured with Infrared (IR) cameras? The research community responded with a continuous effort to find this crucial relationship. This effort was aided by advances in processing techniques, improved sensitivity and spatial resolution of thermal sensors. However, despite this progress fundamental issues with this imaging modality still remain. The main problem is that the link between disease and heat radiation is complex and in many cases even non-linear. Furthermore, the change in heat radiation as well as the change in radiation pattern, which indicate disease, is minute. On a technical level, this poses high requirements on image capturing and processing. On a more abstract level, these problems lead to inter-observer variability and on an even more abstract level they lead to a lack of trust in this imaging modality. In this review, we adopt the position that these problems can only be solved through a strict application of scientific principles and objective performance assessment. Computing machinery is inherently objective; this helps us to apply scientific principles in a transparent way and to assess the performance results. As a consequence, we aim to promote thermography based Computer-Aided Diagnosis (CAD) systems. Another benefit of CAD systems comes from the fact that the diagnostic accuracy is linked to the capability of the computing machinery and, in general, computers become ever more potent. We predict that a pervasive application of computers and networking technology in medicine will help us to overcome the shortcomings of any single imaging modality and this will pave the way for integrated health care systems which maximize the quality of patient care.