Thermographic assessment of skin prick tests in comparison with the routine evaluation methods

Deep Learning Method Applied to Autonomous Image Diagnosis for Prick Test

BACKGROUND

The skin prick test (SPT) is used to diagnose sensitization to antigens. This study proposes a deep learning approach to infer wheal dimensions, aiming to reduce dependence on human interpretation.

METHODS

A dataset of SPT images (n = 5844) was used to infer a convolutional neural network for wheal segmentation (ML model). Three methods for inferring wheal dimensions were evaluated: the ML model; the standard protocol (MA1); and approximation of the area as an ellipse using diameters measured by an allergist (MA2). The results were compared with assisted image segmentation (AIS), the most accurate method. Bland-Altman analysis, distribution analyses, and correlation tests were applied to compare the methods. This study also compared the percentage deviation among these methods in determining the area of wheals with regular geometric shapes (n = 150) and with irregular shapes (n = 150).

RESULTS

The Bland-Altman analysis showed that the difference between methods was not correlated with the absolute area. The ML model achieved a segmentation accuracy of 85.88% and a strong correlation with the AIS method (ρ = 0.88), outperforming all other methods. Additionally, MA1 showed significant error (13.44 ± 13.95%) for pseudopods.

CONCLUSIONS

The ML protocol can potentially automate the reading of SPT, offering greater accuracy than the standard protocol.

Thermography based skin allergic reaction recognition by convolutional neural networks

In this work we present an automated approach to allergy recognition based on neural networks. Allergic reaction classification is an important task in modern medicine. Currently it is done by humans, which has obvious drawbacks, such as subjectivity in the process. We propose an automated method to classify prick allergic reactions using correlated visible-spectrum and thermal images of a patient’s forearm. We test our model on a real-life dataset of 100 patients (1584 separate allergen injections). Our solution yields good results—0.98 ROC AUC; 0.97 AP; 93.6% accuracy. Additionally, we present a method to segment separate allergen injection areas from the image of the patient’s forearm (multiple injections per forearm). The proposed approach can possibly reduce the time of an examination, while taking into consideration more information than possible by human staff.

Objective evaluation of immediate reading skin prick test applying image planimetric and reaction thermometry analyses

The skin prick test is used to diagnose patients' sensitization to antigens through a mediated IgE response. It is a practical and quick exam, but its diagnosis depends on instruments for measuring the allergic response and observer's interpretation. The conventional method for inferring about the allergic reaction is performed from the dimensions of the wheals, which are measured using a ruler or a caliper. To make this diagnosis less dependent on human interpretation, the present study proposes two alternative methods to infer about the allergic reaction: computational determination of the wheal area and a study of the temperature variation of the patient's skin in the puncture region. For this purpose, prick test using histamine was performed on 20 patients randomly selected. The areas were determined by the conventional method using the dimensions of the wheals measured with a digital caliper 30 min after the puncture. The wheal areas were also determined by a Python algorithm using photographs of the puncture region obtained with a smartphone. A variable named circularity deviation was also determined for each analyzed wheal. The temperature variation was monitored using an infrared temperature sensor, which collected temperature data for 30 min. All results were statistically compared or correlated. The results showed that the computational method to infer the wheal areas did not differ significantly from the areas determined by the conventional method (p-value = 0.07585). Temperature monitoring revealed that there was a consistent temperature increase in the first minutes after the puncture, followed by stabilization, so that the data could be adjusted by a logistic equation (R² = 0.96). This adjustment showed that the optimal time to measure the temperature is 800 s after the puncture, when the temperature stabilization occurs. The results have also shown that this temperature stabilization has a significant positive correlation with wheal area (p-value = 0.0015). Thus, we concluded that the proposed computational method is more accurate to infer the wheal area when compared to the traditional method, and that the temperature may be used as an alternative parameter to infer about the allergic reaction.