Significance: Accurate early diagnosis of malignant skin lesions is critical in providing adequate and timely treatment; unfortunately, initial clinical evaluation of similar-looking benign and malignant skin lesions can result in missed diagnosis of malignant lesions and unnecessary biopsy of benign ones.
Aim: To develop and validate a label-free and objective image-guided strategy for the clinical evaluation of suspicious pigmented skin lesions based on multispectral autofluorescence lifetime imaging (maFLIM) dermoscopy.
Approach: We tested the hypothesis that maFLIM-derived autofluorescence global features can be used in machine-learning (ML) models to discriminate malignant from benign pigmented skin lesions. Clinical widefield maFLIM dermoscopy imaging of 41 benign and 19 malignant pigmented skin lesions from 30 patients were acquired prior to tissue biopsy sampling. Three different pools of global image-level maFLIM features were extracted: multispectral intensity, time-domain biexponential, and frequency-domain phasor features. The classification potential of each feature pool to discriminate benign versus malignant pigmented skin lesions was evaluated by training quadratic discriminant analysis (QDA) classification models and applying a leave-one-patient-out cross-validation strategy.
Results: Classification performance estimates obtained after unbiased feature selection were as follows: 68% sensitivity and 80% specificity with the phasor feature pool, 84% sensitivity, and 71% specificity with the biexponential feature pool, and 84% sensitivity and 32% specificity with the intensity feature pool. Ensemble combinations of QDA models trained with phasor and biexponential features yielded sensitivity of 84% and specificity of 90%, outperforming all other models considered.
Conclusions: Simple classification ML models based on time-resolved (biexponential and phasor) autofluorescence global features extracted from maFLIM dermoscopy images have the potential to provide objective discrimination of malignant from benign pigmented lesions. ML-assisted maFLIM dermoscopy could potentially assist with the clinical evaluation of suspicious lesions and the identification of those patients benefiting the most from biopsy examination.
Melanoma is the most aggressive type of skin cancer with an estimated 106,110 new cases in the US in 2021. The 5-year survival rate of patients with early-stage melanoma is ~99%; however, ~13% of melanoma patients are diagnosed with lesions already at intermediate or advance stages, associated with a 5-year survival rate of ~66% and ~27% respectively. The current diagnosis technique involving visual inspection and biopsy often fail to visually distinguish clinically similar lesions; in particular, melanoma can be mistaken for benign lesion pigmented seborrheic keratosis (pSK). In this work, a deep learning model using Long Short-Term Memory (LSTM) networks is trained on the multispectral autofluorescence lifetime dermoscopy images collected from 41 benign lesions including solar lentigo and pSK, and 19 malignant lesions including melanoma, superficial basal cell carcinoma (BCC) and nodular BCC. The model is trained on the image pixels containing concatenated fluorescent decay signals from three emission channels. The posterior probabilities predicted for each pixel location, is used to construct probability maps of the images. Receiver Operator Characteristics (ROC) constructed on the threshold of the median value of the posterior probability map determines the effectiveness in distinguishing benign and malignant lesions. The entire dataset is split into training, validation, and test sets. The hyperparameters are tuned using the validation set while the model performance is estimated using the test set. The mean and standard deviation of the Areas Under the Curve (AUC) of the ROCs generated with 10 random test sets is 0.82 ± 0.04.
Every year more than 5.4 million new cases of skin cancer are reported in the US. Melanoma is the most lethal type with only 5% occurrence rate, but accounts for over 75% of all skin cancer deaths. Non-melanoma skin cancer, especially basal cell carcinoma (BCC) is the most commonly occurring and often curable type that affects more than 3 million people and causes about 2000 deaths in the US annually. The current diagnosis involves visual inspection, followed by biopsy of the lesions. The major drawbacks of this practice include difficulty in border detection causing incomplete treatment and, the inability to distinguish between clinically similar lesions. Melanoma is often mistaken for the benign lesion pigmented seborrheic keratosis (pSK), making it extremely important to differentiate benign and malignant lesions. In this work, a novel feature extraction algorithm based on phasors was performed on the Fluorescence Lifetime Imaging (FLIM) images of the skin to reliably distinguish between benign and malignant lesions. This approach, unlike the standard FLIM data processing method that requires time-deconvolution of the instrument response from the measured time-resolved fluorescence signal, is computationally much simpler and provides a unique set of features for classification. Subsequently, FLIM derived features were selected using a double step cross validation approach that assesses the reliability and the performance of the resultant trained classifier. Promising FLIM-based classification performance was attained for detecting benign from malignant pigmented (sensitivity: ~80%, specificity: 79%, overall accuracy: ~79%) and nonpigmented (sensitivity: ~88%, specificity: 83%, overall accuracy: ~87%) lesions.
Glasses are important materials for novel technologies, as their properties can be tailored by doping and compositional changes. Furthermore, glasses can also be microstructured, making them interesting for optical and photonic applications. Corning Gorilla Glass is an alkali aluminosilicate glass commonly used as protective layer in smart phones and tablets thanks to its outstanding mechanical properties. Recently, it has been demonstrated the use of femtosecond direct laser writing of waveguides in Gorilla Glass, prompting it for integrated photonic/electronic devices. Therefore, it is important to study the nonlinear optical properties of Gorilla Glass as well as their laser-inscribed waveguides, since the effects of the laser writing process on the nonlinearity are not totally understood.
Here we investigate the third-order nonlinear optical properties of waveguides fs-pulses written waveguides in Gorilla Glass, by using the Dispersive-scan (D-scan) method. The nonlinear refractive index measured in the waveguide is lower than the one for the pristine material and its value depends on the writing pulse energy. For waveguides fabricated with pulse energy of 250 nJ, for instance, n2 is about three times lower than the one for the pristine sample. Micro Raman measurements were performed in the microstructured material in order to better understand the mechanisms of laser modification. Raman spectroscopy revealed the reduction and broadening of the high-frequency band related to non-bridging oxygens, which can explain the decrease of n2. Therefore, our results not only show the potential of using D-scan for waveguides nonlinear characterization, but also demonstrate and interpret the decrease of the nonlinear index of refraction in fs-laser micromachined waveguides in Gorilla Glass, which potential implications for photonic devices.
Skin lesions are the most common human cancer diseases, usually, is it diagnosed by clinical visual inspections followed by biopsy. Early detection of these diseases is critical, depending on an accurate and trained dermatologist and can increase the survival rate. Aiming for screening and early diagnose skin lesions many techniques are presented, however, optical techniques are highlighted since they are fast and noninvasive. In this context, fluorescence steady-state and lifetime imaging show potential by being able to image metabolic changes using endogenous contrast. Here it is demonstrated an in vivo label-free multispectral fluorescence lifetime imaging system to distinguish between two types of clinically similar lesions. A pulsed Nd:YAG laser emitting at 355 nm is used to excite the endogenous fluorophores and three channels of acquisition bands are used to imaging the skin. Preliminary results showed differences in the fluorescence lifetime between Bowen and Actinic Keratosis as well as the lesion and the skin around, demonstrating a potential tool to identify the lesion and its edges.
For a precise characterization of time-domain fluorescence lifetime imaging microscopy (FLIM) datasets, an initial processing step is needed to identify the fluorescent impulse response (FIR) at each spatial point in the sample. Hence departing from the measured fluorescent decays, the FIRs are estimated by using the instrument response function (IRF), and this processing step is known as deconvolution. However, the deconvolution methodology requires an initial measurement of the IRF and a corresponding synchronization step with the fluorescent decays. In this context, we propose a blind deconvolution strategy that estimates jointly the FIRs and the IRF in the dataset. For this purpose, each FIR is modeled by a multi-exponential structure. In this way, the FIRs are characterized by the scaling coefficients and time constants of the exponential terms. Meanwhile, there is no explicit model or pre-defined shape for the IRF. Overall estimation process is achieved by an alternated least squares methodology between the FIRs and IRF. First, if the IRF is fixed, a nonlinear least squares framework computes the FIRs parameters at each spatial point of the sample. Meanwhile, once the FIRs are fixed, the samples of the IRF are estimated by a non-negative least squares methodology and using the whole dataset. These alternated optimization steps are performed until a convergence criterion is fulfilled. The proposed blind deconvolution strategy was validated by synthetic datasets and in vivo FLIM oral mucosa measurements. In these tests, our proposal shows good characterizations of the FIRs and the IRFs in the FLIM datasets.
Usually, tissue images at cellular level need biopsies to be done. Considering this, diagnostic devices, such as microendoscopes, have been developed with the purpose of do not be invasive. This study goal is the development of a dual-channel microendoscope, using two fluorescent labels: proflavine and protoporphyrin IX (PpIX), both approved by Food and Drug Administration. This system, with the potential to perform a microscopic diagnosis and to monitor a photodynamic therapy (PDT) session, uses a halogen lamp and an image fiber bundle to perform subcellular image. Proflavine fluorescence indicates the nuclei of the cell, which is the reference for PpIX localization on image tissue. Preliminary results indicate the efficacy of this optical technique to detect abnormal tissues and to improve the PDT dosimetry. This was the first time, up to our knowledge, that PpIX fluorescence was microscopically observed in vivo, in real time, combined to other fluorescent marker (Proflavine), which allowed to simultaneously observe the spatial localization of the PpIX in the mucosal tissue. We believe this system is very promising tool to monitor PDT in mucosa as it happens. Further experiments have to be performed in order to validate the system for PDT monitoring.
Fluorescence spectroscopy and lifetime techniques are potential methods for optical diagnosis and characterization of biological tissues with an in-situ, fast, and noninvasive interrogation. Several diseases may be diagnosed due to differences in the fluorescence spectra of targeted fluorophores, when, these spectra are similar, considering steady-state fluorescence, others may be detected by monitoring their fluorescence lifetime. Despite this complementarity, most of the current fluorescence lifetime systems are not robust and portable, and not being feasible for clinical applications. We describe the assembly of a fluorescence lifetime spectroscopy system in a suitcase, its characterization, and validation with clinical measurements of skin lesions. The assembled system is all encased and robust, maintaining its mechanical, electrical, and optical stability during transportation, and is feasible for clinical measurements. The instrument response function measured was about 300 ps, and the system is properly calibrated. At the clinical study, the system showed to be reliable, and the achieved spectroscopy results support its potential use as an auxiliary tool for skin diagnostics.
The optical microscopy is one of the most powerful tool in the analysis of biological systems. The usual transmitted light microscope uses a white light lamp as source, what sometimes does not bring optimal results, making it necessary to introduce filters to change some illumination properties like the color temperature or the color itself. There is, of course, an intrinsic limitation on the use of filters that is the lack of an analogical control on the illumination properties and a practical limitation that depends on the number of available filters. To address this need, we developed an illumination system based on (Red, Green and Blue) RGB LEDs, were the microscope operator can control the intensity of each one independently and manually. This paper details the developed system and describes the methods used to compare quantitatively the images acquired while using the standard white light illumination and the images obtained with the developed system. To quantify the contrast, we calculated the relative population standard deviation for the intensities of each channel of the RGB image. This procedure allowed us to compare and understand the major advantages of the developed illumination system. All analysis methods have shown that a contrast enhancement can be obtained under the RGB LEDs light. The presented illumination allowed us to visualize the structures in different samples with a better contrast without the need of any additional optical filters.
A portable microscope/microendoscope will be presented in this article. The system was specially designed for Smartphones and taking into account its simplicity, will be able to bring this technology to almost every doctor’s office. It is worth mentioning its flexibility of use, that allows several modes since all the components are interchangeable (the illumination LED, the lens, the optic filters, etc) resulting in different applications, from medical applications until other areas (for example, the inspection of non-accessible pieces of plane engines). In addition, the system has a double platform, working as a conventional microscope or as a fiberoptic microendoscope. In situ and cell smear interrogation of oral mucosa, using a proflavine as dye will be presented. The price of the system does not exceed US$ 350, plus the price of the fiber bundle (around US$ 500) turning it onto a high resolution affordable system.
The fluorescence spectra and fluorescence lifetime analysis in biological tissues has been presented as a technique of a great potential for tissue characterization for diagnostic purposes. The objective of this study is to assemble and characterize a fluorescence lifetime spectroscopy system for diagnostic of clinically similar skin lesions in vivo. The fluorescence lifetime measurements were performed using the Time Correlated Single Photon Counting (Becker & Hickl, Berlin, Germany) technique. Two lasers, one emitting at 378 nm and another at 445 nm, are used for excitation with 20, 50 and 80 MHz repetition rate. A bifurcated optical fiber probe conducts the excitation light to the sample, the collected light is transmitted through bandpass filters and delivered to a hybrid photomultiplier tube detector. The fluorescence spectra were obtained by using a portable spectrometer (Ocean Optics USB-2000-FLG) with the same excitation sources. An instrument response function of about 300 ps was obtained and the spectrum and fluorescence lifetime of a standard fluorescent molecule (Rhodamine 6G) was measured for the calibration of the system ((4.1 ± 0.3) ns). The assembled system was considered robust, well calibrated and will be used for clinical measurements of skin lesions.
The collagen fibers are one of the most important structural proteins in skin, being responsible for its strength and flexibility. It is known that their properties, like fibers density, ordination and mean diameter can be affected by several skin conditions, what makes these properties a good parameter to be used on the diagnosis and evaluation of skin aging, cancer, healing, among other conditions. There is, however, a need for methods capable of analyzing quantitatively the organization patterns of these fibers. To address this need, we developed a method based on the autocorrelation function of the images that allows the construction of vector field plots of the fibers directions and does not require any kind of curve fitting or optimization. The analyzed images were obtained through Second Harmonic Generation Imaging Microscopy. This paper presents a concise review on the autocorrelation function and some of its applications to image processing, details the developed method and the results obtained through the analysis of hystopathological slides of landrace porcine skin. The method has high accuracy on the determination of the fibers direction and presents high performance. We look forward to perform further studies keeping track of different skin conditions over time.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.