Preliminary investigation of skin conditions by visual examination can become highly subjective for doctors, especially when there is a need to accurately differentiate between closely resembling skin conditions. In such ambiguities, invasive procedures like biopsies are employed to confirm the diagnosis.

 

In India around 6% to 11% of the population is affected by skin conditions, especially those living in rural areas which lack the access to proper healthcare facilities. Adding to this, the ratio of patients to dermatologists in India is roughly 200000:1. Most of these dermatologists practice in urban areas, which results in the majority of rural population being examined by non-dermatologists. In such a setup, performing invasive procedures such as biopsy can be very time-consuming, painful and sometimes unnecessarily expensive for the patient. This creates a need for diagnostic aids that can help doctors/primary healthcare workers in preliminary screening of a large number of patients in a quick and non-invasive manner.

 

Fluorescence spectroscopy has been used in various applications related to medical uses, food testing, clinical diagnosis etc. due to its high sensitivity and specificity. When used for skin conditions, studies have shown that the underlying changes in the constituents of the skin leads to generation of condition-specific auto fluorescence response, especially in the case of pigmentary conditions. Attempts have been made to automate the identification and differentiation of skin conditions by using image processing and machine learning tools on dermoscopic images of skin lesions which depend on the external appearance of the skin.

 

There is a need for a method that can help in the non-invasive identification of closely resembling skin conditions such that it can be used for a more effective preliminary screening. The existing optical imaging techniques get affected by ambient light noise. A more controlled and quantified technique such as fluorescence spectroscopy can be more helpful as it is also based on the underlying constitutional changes of the skin. Fluorescence spectroscopy measures the emitted spectra from various skin lesions and helps in identifying physiological changes in different skin disorders.

 

Analysis of the spectral graphs obtained, would determine the utility of the device in differentiating closely resembling skin disorders. The device would be helpful in rapid screening of specific skin disorders and would be very useful in rural and semi-urban areas where social stigma against diseases like vitiligo is very high; especially in the setting of non-availability of dermatologists. We have demonstrated a novel approach that can aid in preliminary screening of skin conditions which can be very useful in developing countries like India where there is a shortage of dermatologists.

 

It is a system comprising of a portable, handheld mini-spectrometer used in conjunction with a smartphone to perform a fluorescence spectroscopy test.

 

The data which is collected by this technique can be used to develop data-driven algorithms to identify and differentiate between various similar looking conditions. This hardware combination is portable and affordable compared to the bulky spectroscopic laboratory setups, making it suitable to be widely used in rural areas of India.