Raman spectroscopy (RS) has demonstrated great potential for in vivo cancer screening; however, the biophysical changes that occur for specific diagnoses remain unclear. We recently developed an inverse biophysical skin cancer model to address this issue. Here, we presented the first demonstration of in vivo melanoma and nonmelanoma skin cancer (NMSC) detection based on this model. We fit the model to our previous clinical dataset and extracted the concentration of eight Raman active components in 100 lesions in 65 patients diagnosed with malignant melanoma (MM), dysplastic nevi (DN), basal cell carcinoma, squamous cell carcinoma, and actinic keratosis. We then used logistic regression and leave-one-lesion-out cross validation to determine the diagnostically relevant model components. Our results showed that the biophysical model captures the diagnostic power of the previously used statistical classification model while also providing the skin’s biophysical composition. In addition, collagen and triolein were the most relevant biomarkers to represent the spectral variances between MM and DN, and between NMSC and normal tissue. Our work demonstrates the ability of RS to reveal the biophysical basis for accurate diagnosis of different skin cancers, which may eventually lead to a reduction in the number of unnecessary excisional skin biopsies performed.
The recurrence rate of nonmelanoma skin cancer is highly related to the residual tumor after surgery. Although tissueconserving surgery, such as Mohs surgery, is a standard method for the treatment of nonmelanoma skin cancer, they are limited by lengthy and costly frozen-section histopathology. Raman spectroscopy (RS) is proving to be an objective, sensitive, and non-destructive tool for detecting skin cancer. Previous studies demonstrated the high sensitivity of RS in detecting tumor margins of basal cell carcinoma (BCC). However, those studies rely on statistical classification models and do not elucidate the skin biophysical composition. As a result, we aim to discover the biophysical differences between BCC and primary normal skin structures (including epidermis, dermis, hair follicle, sebaceous gland and fat). We obtained freshly resected ex vivo skin samples from fresh resection specimens from 14 patients undergoing Mohs surgery. Raman images were acquired from regions containing one or more structures using a custom built 830nm confocal Raman microscope. The spectra were grouped using K-means clustering analysis and annotated as either BCC or each of the five normal structures by comparing with the histopathology image of the serial section. The spectral data were then fit by a previously established biophysical model with eight primary skin constituents. Our results show that BCC has significant differences in the fit coefficients of nucleus, collagen, triolein, keratin and elastin compared with normal structures. Our study reveals RS has the potential to detect biophysical changes in resection margins, and supports the development of diagnostic algorithms for future intraoperative implementation of RS during Mohs surgery.
Mohs surgery is the current gold standard to treat large, aggressive or high-risk non-melanoma skin cancer (NMSC) cases. While Mohs surgery is an effective treatment, the procedure is time-consuming and expensive for physicians as well as burdensome for patients as they wait for frozen section histology. Our group has recently demonstrated high diagnostic accuracy using a noninvasive “spectral biopsy” (combination of diffuse reflectance (DRS), fluorescence (FS) and Raman spectroscopy (RS)) to classify NMSC vs. normal lesion in a screening setting of intact tissue. Here, we examine the sensitivity of spectral biopsy to pathology in excised Mohs sections. The system is designed with three modalities integrated into one fiber probe, which is utilized to measure DRS, FS, and RS of freshly excised skin from patients with various NMSC pathologies including basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), where each measurement location is correlated to histopathology. The spectral biopsy provides complimentary physiological information including the reduced scattering coefficient, hemoglobin content and oxygen saturation from DRS, NADH and collagen contribution from FS and information regarding multiple proteins and lipids from RS. We then apply logistic regression model to the extracted physiological parameters to classify NMSC vs. normal tissue. The results on the excised tissue are generally consistent with in vivo measurements showing decreased scattering within the tumor and reduced fluorescence. Due to the high sensitivity of RS to lipids, subcutaneous fat often dominates the RS signal. This pilot study demonstrates the potential for a spectral biopsy to classify NMSC vs. normal tissue, indicating the opportunity to guide Mohs excisions.
Skin cancer is the most common form of cancer in the United States and is a recognized public health issue. Diagnosis of
skin cancer involves biopsy of the suspicious lesion followed by histopathology. Biopsies, which involve excision of the
lesion, are invasive, at times unnecessary, and are costly procedures (~$2.8B/year in the US). An unmet critical need
exists to develop a non-invasive and inexpensive screening method that can eliminate the need for unnecessary biopsies.
To address this need, our group has reported on the continued development of a noninvasive method that utilizes
multimodal spectroscopy towards the goal of a “spectral biopsy” of skin. Our approach combines Raman spectroscopy,
fluorescence spectroscopy, and diffuse reflectance spectroscopy to collect comprehensive optical property information
from suspicious skin lesions. We previously described an updated spectral biopsy system that allows acquisition of all
three forms of spectroscopy through a single fiber optic probe and is composed of off-the-shelf OEM components that
are smaller, cheaper, and enable a more clinic-friendly system. We present initial patient data acquired with the spectral
biopsy system, the first from an extensive clinical study (n = 250) to characterize its performance in identifying skin
cancers (basal cell carcinoma, squamous cell carcinoma, and melanoma). We also present our first attempts at analyzing
this initial set of clinical data using statistical-based models, and with models currently being developed to extract
biophysical information from the collected spectra, all towards the goal of noninvasive skin cancer diagnosis.