Incidence and profile of skin cancers in patients following ultraviolet phototherapy without psoralens: A retrospective cohort study.

BACKGROUND: Psoralen + ultraviolet-A (PUVA) is associated with photocarcinogenesis. However, carcinogenic risk with other ultraviolet phototherapies remains unclear. OBJECTIVE: Evaluate whether phototherapy without psoralens increases skin cancer risk. METHODS: Retrospective cohort study of patients treated at a teaching-hospital phototherapy center (1977-2018). Skin cancer records were validated against pathology reports. Age-standardized incidence rates (ASIRs) of skin cancer were evaluated for gender, skin phototype, diagnosis, ultraviolet modality, anatomical site; and compared to provincial population incidence rates (2003). RESULTS: In total, 3506 patients treated with broadband-ultraviolet-B, narrowband-UVB and/or combined UVAB were assessed with a mean follow-up of 7.3 years. Majority of patients had psoriasis (60.9%) or eczema (26.4%). Median number of treatments was 43 (1-3598). Overall, 170 skin cancers (17 melanoma, 33 squamous cell carcinoma and 120 basal cell carcinoma) occurred in 79 patients. Patient-based and tumor-based ASIR of skin cancer was 149 (95% CI: 112-187)/100,000 and 264 (219-309)/100,000 person-years, respectively. There was no significant difference between tumor-based ASIRs for melanoma, squamous cell carcinoma, and basal cell carcinoma compared to the general population; or in phototherapy patients with-psoriasis or eczema; or immunosuppressants. No cumulative dose-response correlation between UVB and skin cancer was seen. LIMITATIONS: Treatment and follow-up duration. CONCLUSION: No increased risk of melanoma and keratinocyte cancer was found with phototherapy.

Baseline Characteristics of Canadian Patients in the Psoriasis Longitudinal Assessment and Registry (PSOLAR).

BACKGROUND: The Psoriasis Longitudinal Assessment and Registry (PSOLAR) is a global, prospective, longitudinal, disease-based registry. It serves as a post-marketing safety commitment with a focus on patients with moderate to severe plaque psoriasis who are candidates for systemic therapy. OBJECTIVES: To describe the baseline disease demographics and clinical characteristics of a Canadian subgroup of participants enrolled in PSOLAR. METHODS: Baseline demographic/disease characteristics, medical histories, and previous psoriasis treatments for Canadian patients in PSOLAR were summarized using descriptive statistics. RESULTS: There were 1896 patients analyzed in the Canadian subgroup at 37 clinical sites, accounting for 15.7% of the global PSOLAR population. Baseline disease and clinical characteristics were as expected for a moderate to severe psoriasis population and were generally similar to the global PSOLAR population. Two distinctions were noted in the Canadian subgroup versus those enrolled globally: a higher proportion of patients were overweight/obese (84.7% vs. 80.4%) and male (61.4% vs. 54.7%). In addition, the Canadian subgroup had numerically higher historical peak disease activity (PGA score 3.35 vs. 3.1) and longer disease duration (22.3 years vs. 17.5 years). Canadian PSOLAR patients reported a variety of comorbidities, including psoriatic arthritis (31.5%), hypertension (34.6%), hyperlipidemia (24.3%), mental illness (24.1%), and inflammatory bowel disease (1.6%). CONCLUSION: The Canadian subgroup of PSOLAR patients was generally similar to those enrolled globally with respect to baseline disease demographics and clinical characteristics. Multiple comorbidities are noted in the Canadian subgroup, underscoring the need for a holistic approach to the treatment of psoriatic patients.

Assessment of sun-safety education behavior via spectrophotometric evaluation: A preliminary study.

BACKGROUND: Biases inherent in self-assessment of sun exposure and sun-safe behavior may lead to inaccurate conclusions about the effectiveness of sun-safety educational programs. OBJECTIVES: We aimed to compare self-reports to objective measures of sun exposure, when examining the effectiveness of passive versus active educational interventions. METHODS: From May to June 2018, 73 participants recruited at a dermatology clinic were sequentially assigned to receive sun-safety education through one of 3 modes: interactive online module, video, or no education. A baseline Sun Exposure and Behavior Inventory (SEBI) questionnaire was administered, and spectrophotometric measurements of sun-exposed and sun-protected areas were taken and reported in the CIE L*a*b* color space. Participants were followed 4-8 and 16 weeks after the initial visit where the SEBI was re-administered, and serial measurements of skin color were taken. The change in SEBI scores and L*a*b values, as calculated by the individual typology angle (ITA°), was analyzed. RESULTS: There was a significant increase in skin darkening in all the groups at 4-8 and 16 weeks follow-up. There was no statistically significant difference between the groups in the magnitude of color change. However, subjectively at 4-8 weeks post-intervention, participants in the interactive module and video groups had significantly improved self-reported SEBI scores compared to control (p < .05, Kruskal-Wallis). By 16 weeks, only the interactive module group showed significant improvement in SEBI scores compared to control (p < .05, ANOVA). CONCLUSION: In determining the effectiveness of sun-safety programs, spectrophotometric evaluation of sun-induced skin pigmentation can allow for a more complete evaluation of self-reported sun exposure and sun-protective behavior.

Solar ultraviolet radiation exposure among outdoor workers in Alberta, Canada.

BACKGROUND: Outdoor workers are at risk of prolonged and high solar ultraviolet radiation (UVR) exposure, which is known to cause skin cancer. The objectives of this study were to characterize the UVR exposure levels of outdoor workers in Alberta, Canada, and to investigate what factors may contribute to their exposure. METHODS: This study collected objective solar UVR measurements from outdoor workers primarily in Alberta during the summer of 2019. Workers were recruited via the management or health and safety teams from building trade unions and employers. Calibrated, electronic UVR dosimeters were worn by workers on their hardhats, wrists, or lapels for five working days. Data on workers' demographics, jobs, sun protection behaviors, and personal risk factors were collected using questionnaires, and meteorological data for each sampling day were noted. Mean daily exposure measured as the standard erythemal dose (SED) was calculated and compared to the international occupational exposure limit guideline (1.3 SED). Marginal models were developed to evaluate potential determinants of occupational solar UVR exposure. RESULTS: In total, 883 measurements were collected from 179 workers. On average, workerswere exposed to 1.93 SED (range: 0.03-16.63 SED) per day. Just under half of workers (45%) were exposed to levels exceeding the international exposure limit guideline. In the bivariate analyses, landscape and maintenance workers, as well as trade and recreation workers, had the highest levels of exposure (average: 2.64 and 1.84 SED, respectively). Regional variations were observed, with the "other" cities/regions (outside of Edmonton and Calgary) experiencing the highest average levels (2.60 SED). Workers who placed the dosimeters on their hardhats experienced higher levels compared to the other groups. Exposure was highest on sunny and mixed days. Education, trade, city, dosimeter placement, forecast, hair colour, and number of hours outside were included in the final exposure model, of which trade, dosimeter placement, forecast, and number of hours outside at work were statistically significant. CONCLUSIONS: Exposure to elevated solar UVR levels is common among outdoor workers in Alberta. The study findings can help inform future monitoring studies and exposure reduction initiatives aimed at protecting workers.

Relationship between sun safety behaviours and modifiable lifestyle cancer risk factors and vitamin D levels.

BACKGROUND: Sun exposure is the most important environmental risk factor for causing skin cancer. PURPOSE: This study examines the relationship between sun protection behaviours and modifiable lifestyle risk factors for other cancers as well as vitamin D levels. METHODS: Cross-sectional data were analysed from two large national health surveys (n = 31, 445 and n = 5604). Sun exposure and protection were characterized by the presence of sunburn, duration of sun exposure, frequency of seeking shade, frequency of wearing a hat and frequency of wearing sunscreen. Using Statistical Analysis System (SAS) software 9.3.1, multivariate logistic regression models were compiled. RESULTS: Unhealthy behaviour practices were associated with sunburns or infrequent sun protection behaviour, such as cigarette consumption (either current or former smokers), second-hand smoke exposure, not having a regular doctor, higher level of alcohol consumption, street drug usage and low levels of fruit/vegetable consumption. Approximately one-quarter of individuals had less than the recommended value of serum vitamin D levels (<50 nmol/L), despite 39.2% of these individuals reporting ≥1 hour of sun exposure. CONCLUSION: Modifiable lifestyle risk factors for other cancers are correlated with infrequently practicing sun protection behaviours for skin cancer prevention. Therefore, cancer prevention campaigns can aim to target all these risk factors associated with different cancers. Sun exposure is not a reliable source to obtain recommended vitamin D levels and that other sources (eg. fish, egg yolk, fortified drinks and supplements) are a safer and more reliable option.

Site-specific variations in cutaneous autofluorescence revealed by excitation-emission matrix spectroscopy.

BACKGROUND AND PURPOSE: Although cutaneous autofluorescence has been utilized for evaluation of skin conditions, there is a paucity of data on normal human skin autofluorescence and its dependence on anatomical site. The objective of this study is to use excitation-emission matrix spectroscopy to quantify and characterize skin autofluorescence at different body sites. METHODS: Ten anatomical sites from 30 healthy volunteers were measured with a double-grating excitation-emission matrix spectrofluorometer. RESULTS: For the 10 body sites evaluated, there were four overall patterns of autofluorescence: skin from the head and neck exhibits high superficial and low bilayer fluorescence; the dorsal forearm and dorsal hand have both low superficial and bilayer fluorescence; the upper inner arm and back have high superficial and intermediate bilayer fluorescence; while the palm and thumbnail have both high superficial and bilayer fluorescence. The corresponding fluorescence excitation-emission peaks for these patterns were as follows: head and neck, 3 peaks at 290-300/330-350, 360-380/460-485, and 380-420/610-630 nm; dorsal forearm and dorsal hand, 2 peaks around 295-300/345-360 and 385-395/460-485 nm; upper inner arm and back, 3 peaks around 295-300/335-355, 335-340/390-410, and 375-390/455-480 nm; palm and thumbnail, 3 peaks around 285-300/345-355, 335-345/390-410, and 365-390/450-480 nm. CONCLUSION: Cutaneous fluorescence varies in distinct patterns according to anatomical site, due to the component fluorophores present, skin thickness, and the degree of melanization and long term sun exposure. These EEM patterns for normal skin should be accounted for when interpreting fluorescence signals from disease states and can also be used to guide the selection of optimal wavebands when applying this optical modality.

Non-Melanoma Skin Cancer Risk Among Patients in the Psoriasis Longitudinal Assessment and Registry (PSOLAR)

To the Editor: Patients with psoriasis are at increased risk of developing non melanoma skin cancer (NMSC), including squamous cell carcinoma (SCC) and basal cell carcinoma (BCC).1,2 The risk is especially elevated among those who previously received systemic treatment or phototherapy.2 Systemic treatments, including biologic therapies and methotrexate (MTX), are effective in managing immune-mediated diseases; however, they may increase susceptibility to NMSC due to immunosuppression or other factors.3

A Qualitative Analysis of Canadian Indoor Tanning Policies.

BACKGROUND: The International Agency for Research on Cancer classifies artificial tanning devices as Group 1 human carcinogens. Studies have shown that use of indoor tanning before age 35 can increase the risk of melanoma development by 75%. It has therefore been recommended that indoor tanning use be restricted in individuals younger than age 18. OBJECTIVES: This study aims to review the state of provincial indoor tanning policies, especially in regards to use by youth across Canada, and what strategies are being implemented to enforce them. METHODS: Focused interviews were conducted with representatives from the provincial Ministries of Health across Canada in May and June 2014. Follow-up interviews were performed between February and May 2017. RESULTS: As of January 2018, regulations are in effect in all Canadian provinces restricting indoor tanning by minors and requiring display of signage warning of the risks of indoor tanning by salons. However, there are discrepancies among the provinces on how and if tanning salons are monitored and how and if these regulations are enforced. CONCLUSIONS: While implementing youth bans on indoor tanning is a promising start, all Canadian provinces need to ensure that efforts are being undertaken to ensure compliance with these policies to effectively combat the rising incidence of skin cancer among the Canadian population.

Prescribing Home Narrowband UVB Phototherapy: A Review of Current Approaches.

Conventional, full-body phototherapy equipment is costly and therefore patients are usually treated in dermatology centres. Such office-based therapy is often not feasible for those patients who live far away from a phototherapy centre due to lost time and wages, inability to travel because of extensive skin disease, or prohibitive travel costs. Home phototherapy has emerged as a modality that meets the needs of those patients. Our aim was to review available studies on UV sources, treatment protocols, efficacy, and safety of home phototherapy. A literature review was conducted on PubMed using the terms "home" AND "phototherapy" AND ("guide" OR "approach" OR "review" OR "protocol"). From the data extracted, narrowband UVB (311 nm) offers the best balance between safety and efficacy and is recommended for home phototherapy by most authors. Treatment is safe and possible adverse effects are related to overexposure (erythema, blistering). The usual treatment protocol was administering treatments on alternating days, including weekends, with dosing based on the patient's Fitzpatrick skin type. We also provide information on the available home phototherapy systems in Canada and their reimbursement. Home phototherapy is underused in Canada. Narrowband UVB phototherapy sources equipped with a 3-dimensional panel provides a practical and safe option.

Quantifying the visual appearance of sunscreens applied to the skin using indirect computer image colorimetry.

BACKGROUND: There is no accepted method to objectively assess the visual appearance of sunscreens on the skin. METHODS: We present a method for sunscreen application, digital photography, and computer analysis to quantify the appearance of the skin after sunscreen application. Four sunscreen lotions were applied randomly at densities of 0.5, 1.0, 1.5, and 2.0 mg/cm2 to areas of the back of 29 subjects. Each application site had a matched contralateral control area. High-resolution standardized photographs including a color card were taken after sunscreen application. After color balance correction, CIE L*a*b* color values were extracted from paired sites. Differences in skin appearance attributed to sunscreen were represented by ΔE, which in turn was calculated from the linear Euclidean distance within the L*a*b* color space between the paired sites. RESULTS: Sunscreen visibility as measured by median ΔE varied across different products and application densities and ranged between 1.2 and 12.1. The visibility of sunscreens varied according to product SPF, composition (organic vs inorganic), presence of tint, and baseline b* of skin (P < .05 for all). CONCLUSION: Standardized sunscreen application followed by digital photography and indirect computer-based colorimetry represents a potential method to objectively quantify visibility of sunscreen on the skin.

Clinical Insights About Onychomycosis and Its Treatment: A Consensus.

BACKGROUND: Recently, experience and knowledge have been gained using effective topical treatment for onychomycosis, a difficult-to-treat infection. METHODS: This project aims to help understand and improve patient-focused quality of care for fungal nail infections. A panel of dermatologists who treat onychomycosis convened on several occasions to review and discuss recent learnings in the treatment of onychomycosis. The panel developed and conducted a survey on diagnosis, treatment and prevention, discussed the results, and provided recommendations. RESULTS: The survey was sent out digitally to the Canadian Dermatology community. Ninety-two dermatologists completed the questionnaires, which were included in the analysis. The survey respondents and panel members agreed that the diagnosis of toe onychomycosis should be confirmed with a positive microscopic examination for fungus or a positive mycological culture when oral therapy and/or topical treatment is prescribed, except when it is not clinically feasible, in which case topical therapy could be started based on clinical presentation. The panel and survey respondents also agreed that treatment is to be based on percentage of nail involvement: less than 20%=topical efinaconazole; 20%-60%=topical efinaconazole±oral terbinafine (for greater than 3 nails); greater than 60%=oral terbinafine±topical therapy. CONCLUSIONS: The current treatment paradigm for onychomycosis may have shifted from mainly oral antifungals to topical treatment, improving patient-focused quality of care.

J Drugs Dermatol. 2018;17(3):253-262.

Approach to the Management of Patients With Hidradenitis Suppurativa: A Consensus Document.

BACKGROUND: Hidradenitis suppurativa (HS) is a painful, debilitating, and poorly understood condition, which is suboptimally diagnosed, managed, and treated. Evidence supporting various treatment modalities is sparse. OBJECTIVES: To incorporate scientific evidence and expert opinions to develop useful guidance for the evaluation and management of patients with HS. METHODS: An expert panel of Canadian dermatologists and surgeons developed statements and recommendations based on available evidence and clinical experience. The statements and recommendations were subjected to analysis and refinement by the panel, and voting was conducted using a modified Delphi technique with a prespecified cutoff agreement of 75%. RESULTS: Ten specific statements and recommendations were accepted by the expert panel. These were grouped into 4 domains: diagnosis and assessment, treatment and management, comorbidities and a multidisciplinary approach, and education. CONCLUSIONS: These statements and recommendations will serve to increase awareness of HS and provide a framework for decisions involving diagnosis and management. Evidence suggests that antibacterial and anti-tumour necrosis factor therapies are effective in the treatment of HS. This is supported by the clinical experience of the authors. Further clinical research and the establishment of multidisciplinary management teams will continue to advance management of HS in Canada.

Wavenumber selection based analysis in Raman spectroscopy improves skin cancer diagnostic specificity.

Real-time Raman spectroscopy can be used to assist in assessing skin lesions suspicious for cancer. Most of the diagnostic algorithms are based on full band of the Raman spectra, either in the fingerprint region or the high wavenumber region. In this paper we explored wavenumber selection based analysis in Raman spectroscopy for skin cancer diagnosis. Wavenumber selection was implemented using windows of wavenumber and leave-one-out cross-validated stepwise regression or least and shrinkage selection operator (LASSO). The diagnostic algorithms were then generated from the selected windows of wavenumber using multivariate statistical analyses, including principal component and general discriminate analysis (PC-GDA) and partial least squares (PLS). In total a combined cohort of 645 confirmed lesions from 573 patients encompassing skin cancers, precancers and benign skin lesions were included, which were divided into training cohort (n = 518) and testing cohort (n = 127) according to the measurement time. It was found that the area under the receiver operating characteristic curve (ROC) was improved from 0.861-0.891 to 0.891-0.911 and the diagnostic specificity for fixed sensitivity 0.99-0.90 was improved from 0.17-0.65 to 0.20-0.75 with wavenumber selection based analysis.

The prevalence of anxiety and depression in patients with or without hyperhidrosis (HH).

BACKGROUND: There are conflicting data about the correlation between hyperhidrosis (HH) and anxiety and depression. OBJECTIVE: We sought to determine the prevalence of anxiety and depression in patients with or without HH. METHODS: We examined 2017 consecutive dermatology outpatients from Vancouver, British Columbia, Canada, and Shanghai, China, using Patient Health Questionnaire-9 and Generalized Anxiety Disorder-7 scales for anxiety and depression assessments. Multivariable logistic regression analysis was performed to evaluate if the impact of HH on anxiety and depression is dependent on demographic factors and diagnoses of the patients' presenting skin conditions. RESULTS: The prevalence of anxiety and depression was 21.3% and 27.2% in patients with HH, respectively, and 7.5% and 9.7% in patients without HH, respectively (P value <.001 for both). There were positive correlations between HH severity and the prevalence of anxiety and depression. Multivariable analysis showed that HH-associated increase in anxiety and depression prevalence is independent of demographic factors and presenting skin conditions. LIMITATION: The data from the questionnaires relied on the accuracy of patients' self-reports. CONCLUSION: Both single variant and multivariable analyses showed a significant association between HH and the prevalence of anxiety and depression in a HH severity-dependent manner.

Occupational exposure to solar ultraviolet radiation and the risk of prostate cancer.

OBJECTIVES: Preventable risk factors for prostate cancer are poorly understood; sun exposure is a possible protective factor. The goal of this study was to investigate prostate cancer risk in outdoor workers, a population with high sun exposure. METHODS: Prostate cancer cases and controls from a large study (conducted between 1994 and 1997) were used for this analysis. A job exposure matrix (JEM) was used to assign solar ultraviolet radiation (UVR) at work as moderate (2 to <6 hours outside/day) or high (≥6 hours). Average daily satellite UV-B measures were linked to the latitude/longitude of the residences of each participant. Several other exposure metrics were also examined, including ever/never exposed and standard erythemal dose by years (SED×years). Logistic regression was used to evaluate the association between solar UVR exposure and the odds of prostate cancer. RESULTS: A total of 1638 cases and 1697 controls were included. Men of Indian and Asian descent had reduced odds of prostate cancer (ORs 0.17 (0.08 to 0.35) and 0.25 (0.15 to 0.41), respectively) compared with Caucasian men, as did single men (OR 0.76 (0.58 to 0.98)) compared with married men. Overall, no statistically significant associations were observed between sun exposure and prostate cancer with 1 exception. In the satellite-enhanced JEM that considered exposure in high category jobs only, prostate cancer odds in the highest quartile of cumulative exposure was decreased compared with unexposed men (OR 0.68 (0.51 to 0.92)). CONCLUSIONS: This study found limited evidence for an association with prostate cancer, with the exception of 1 statistically significant finding of a decreased risk among workers with the longest term and highest sun exposure.

Levels of Occupational Exposure to Solar Ultraviolet Radiation in Vancouver, Canada.

INTRODUCTION: Outdoor workers are at high risk of exposure to solar ultraviolet radiation (UVR), a known human carcinogen. In Canada, no objective measures of UVR exposure are available for occupational settings. METHODS: The Outdoor Workers Project collected UVR exposure data among outdoor workers in Vancouver, Canada during the summer of 2013. Objective measures of exposure were taken for one week using calibrated electronic UVR dosimeters. Additional data was collected from workers on skin cancer risk factors, family history of skin cancer, and job type; as well as meteorological data for sampling days. Marginal models were constructed to examine the worker, job and meteorological determinants of UVR exposure levels, as measured in standard erythemal dose (SED). RESULTS: Seventy-eight workers were recruited, of which 73 had at least 1 day of measured UVR exposure for this analysis. Participants were mostly male, young and Caucasian. Mean exposure (corrected for repeated measures) was 1.08 SED. Exposure measures were highly variable even in the same workplace, ranging from 0.01 SED to 19.2 SED. Younger age, working in land-based construction, and sunnier weather forecasts led to higher levels of UVR exposure. CONCLUSIONS: Exposure levels capable of causing sunburn were common in this study of outdoor workers, in a location not typically associated with high sun exposure.

Real-time Raman spectroscopy for automatic in vivo skin cancer detection: an independent validation.

In a recent study, we have demonstrated that real-time Raman spectroscopy could be used for skin cancer diagnosis. As a translational study, the objective of this study is to validate previous findings through a completely independent clinical test. In total, 645 confirmed cases were included in the analysis, including a cohort of 518 cases from a previous study, and an independent cohort of 127 new cases. Multi-variant statistical data analyses including principal component with general discriminant analysis (PC-GDA) and partial least squares (PLS) were used separately for lesion classification, which generated similar results. When the previous cohort (n = 518) was used as training and the new cohort (n = 127) was used as testing, the area under the receiver operating characteristic curve (ROC AUC) was found to be 0.889 (95 % CI 0.834-0.944; PLS); when the two cohorts were combined, the ROC AUC was 0.894 (95 % CI 0.870-0.918; PLS) with the narrowest confidence intervals. Both analyses were comparable to the previous findings, where the ROC AUC was 0.896 (95 % CI 0.846-0.946; PLS). The independent study validates that real-time Raman spectroscopy could be used for automatic in vivo skin cancer diagnosis with good accuracy.

Improving skin cancer detection by Raman spectroscopy using convolutional neural networks and data augmentation.

Background: Our previous studies have demonstrated that Raman spectroscopy could be used for skin cancer detection with good sensitivity and specificity. The objective of this study is to determine if skin cancer detection can be further improved by combining deep neural networks and Raman spectroscopy. Patients and methods: Raman spectra of 731 skin lesions were included in this study, containing 340 cancerous and precancerous lesions (melanoma, basal cell carcinoma, squamous cell carcinoma and actinic keratosis) and 391 benign lesions (melanocytic nevus and seborrheic keratosis). One-dimensional convolutional neural networks (1D-CNN) were developed for Raman spectral classification. The stratified samples were divided randomly into training (70%), validation (10%) and test set (20%), and were repeated 56 times using parallel computing. Different data augmentation strategies were implemented for the training dataset, including added random noise, spectral shift, spectral combination and artificially synthesized Raman spectra using one-dimensional generative adversarial networks (1D-GAN). The area under the receiver operating characteristic curve (ROC AUC) was used as a measure of the diagnostic performance. Conventional machine learning approaches, including partial least squares for discriminant analysis (PLS-DA), principal component and linear discriminant analysis (PC-LDA), support vector machine (SVM), and logistic regression (LR) were evaluated for comparison with the same data splitting scheme as the 1D-CNN. Results: The ROC AUC of the test dataset based on the original training spectra were 0.886±0.022 (1D-CNN), 0.870±0.028 (PLS-DA), 0.875±0.033 (PC-LDA), 0.864±0.027 (SVM), and 0.525±0.045 (LR), which were improved to 0.909±0.021 (1D-CNN), 0.899±0.022 (PLS-DA), 0.895±0.022 (PC-LDA), 0.901±0.020 (SVM), and 0.897±0.021 (LR) respectively after augmentation of the training dataset (p<0.0001, Wilcoxon test). Paired analyses of 1D-CNN with conventional machine learning approaches showed that 1D-CNN had a 1-3% improvement (p<0.001, Wilcoxon test). Conclusions: Data augmentation not only improved the performance of both deep neural networks and conventional machine learning techniques by 2-4%, but also improved the performance of the models on spectra with higher noise or spectral shifting. Convolutional neural networks slightly outperformed conventional machine learning approaches for skin cancer detection by Raman spectroscopy.