Overview of Clinical Studies Evaluating DermaSensor
The DermaSensor Instructions for Use (IFU) describes a total of ten clinical studies that provide clinical evidence supporting the safety and effectiveness of the DermaSensor device and algorithm. These studies include early clinical effectiveness studies, principal clinical validation and utility studies demonstrating the device standalone performance and its impact on physician decision-making.
1. DERM-ASSESS II
Study Title: DermaSensor Use in the Assessment of Skin Lesions Suggestive of Cancer (DERM-ASSESS II)
Study Design: U.S.-based, multicenter, prospective, blinded clinical study conducted at four dermatology sites.
Study Population: 333 biopsied skin lesions, including 169 malignant and 164 benign lesions.
Study Objective: The study was conducted to train the device algorithm and validate clinical performance for melanoma, basal cell carcinoma (BCC), and squamous cell carcinoma (SCC).
Results: Device sensitivity was 97%, with a specificity of 26% for biopsied lesions.
Interpretation: The study demonstrated high sensitivity for detecting malignant skin lesions. Overall device specificity was 26%, thus DermaSensor could have effectively ruled out skin cancer for 26% of benign lesions that were biopsied while detecting 97% of skin cancers.
2. Skin Center Investigator-Initiated Study (IIS)
Study Title: Use of an ESS-based Adjunctive Tool in the Assessment of Lesions Suggestive of Skin Cancer
Study Design: Single-site, prospective, blinded clinical study conducted at one dermatology site in New Zealand.
Study Population: 410 skin lesions, including 50 malignant and 360 benign lesions.
Study Objective: The study evaluated DermaSensor device performance as an adjunctive tool for assessing patients with skin lesions suggestive of melanoma, BCC, and SCC.
Results: Device sensitivity was 96.0%, with a specificity of 31.4%.
Interpretation: This study also demonstrated high sensitivity of the DermaSensor device. Device specificity of 31% could have effectively ruled out 31% of benign biopsied lesions while detecting 97% of skin cancers.
3. PATIENT-SELECT Study
Study Title: Use of DermaSensor on Patient-Selected Lesions that are Concerning for Skin Cancer
Study Design: U.S.-based, prospective, blinded clinical study conducted at one primary care site.
Study Population: 178 skin lesions, including 10 malignant and 145 benign lesions.
Study Objective: The study assessed device performance on skin lesions identified by patients themselves as concerning for skin cancer.
Results: Device sensitivity was 90%, with a specificity of 61% (77% for pigmented lesions). The study healthcare professionals’ sensitivity was 40% and specificity 84%. The overall accuracy (i.e. Area Under the Curve or AUC) of the device was greatly higher than that of the study HCPs (0.815 versus 0.643, respectively).
Interpretation: This study evaluated a different lesion-selection approach compared with the HCP-selected lesions investigated in most of the other clinical studies. While the malignant lesion sample size was small, the large benign lesion sample size demonstrated that device specificity is high for lesions of concern to patients enabling physicians to reassure patients that most of their concerning lesions are benign while also potentially improving frontline providers detection of skin cancer.
4. DERM-ASSESS II Utility Study
Study Title: A Multi-Reader, Multi-Case Companion Study to the DERM-ASSESS II Clinical Study
Study Design: U.S.-based, prospective, multi-reader, multi-case study.
Study Population: 57 Primary Care Physicians (PCPs) completed a total of 5,700 lesion assessments.
Study Objective: The study assessed and compared the management sensitivity and specificity of PCPs for melanoma, BCC, and SCC with and without knowledge of the DermaSensor device result.
Results: PCP sensitivity without the device result was 81%. With knowledge of the DermaSensor device result, PCP sensitivity increased to 94%. The difference was statistically significant. Management specificity was 36% without device use and 31% with device use. This difference was not statistically significant.
Interpretation: The study demonstrated that availability of the DermaSensor device result significantly increased PCP sensitivity for malignant skin lesions with no significant decrease in specificity.
5. DERM-SUCCESS Pivotal Validation Study
Study Title: DermaSensor Study of Primary Care Physician Use of Elastic-Scattering Spectroscopy (ESS) on Skin Lesions Suggestive of Skin Cancer
Study Design: International, multicenter, prospective, blinded clinical study conducted at 22 primary care sites, including 18 sites in the United States and four sites in Australia.
Study Population: 1,005 subjects with a total of 1,579 biopsied skin lesions, including 224 malignant lesions.
Study Objective: The study assessed device performance in classifying biopsied lesions suggestive of skin cancer when the device was used by PCPs. Pathology served as the reference standard for all lesions.
Results: DermaSensor device sensitivity was 96%. For comparison, PCP sensitivity was 83%.
Device sensitivity was statistically superior to PCP sensitivity. The study also demonstrated that device performance was non-inferior to the predefined 90% sensitivity performance goal based on dermatologists’ management performance in literature. Device specificity was 21% for benign lesions that were biopsied by the study physicians for suspicion of skin cancer.
Interpretation: Due to its large study population, prospective and multicenter design, and pathology-based reference standard, DERM-SUCCESS represents one of the principal clinical validation studies supporting DermaSensor. The study demonstrated high sensitivity for detecting malignant skin lesions. Overall device specificity was 26%, thus DermaSensor could have effectively ruled out skin cancer for 21% of benign lesions that were biopsied while detecting 96% of skin cancers.
6. DERM-SUCCESS Pivotal Utility Study
Study Title: A Multi-Reader, Multi-Case Companion Study to the DERM-SUCCESS Clinical Study
Study Design: U.S.-based, prospective, multi-reader, multi-case study.
Study Population: 108 board-certified PCP readers completed a total of 10,800 assessments.
Study Objective: The study assessed whether availability of the DermaSensor device result could improve PCP referral decisions and detection of skin cancers.
Results: PCP management sensitivity without the DermaSensor device result was 82%. With knowledge of the device result, PCP sensitivity increased to 91% with a non-significant decrease in management specificity. Overall diagnostic performance, measured by the area under the receiver operating characteristic curve (AUROC), increased from 0.708 to 0.762.
Interpretation: The study demonstrated that availability of the DermaSensor device result can improve PCP sensitivity and overall performance in assessing lesions suggestive of skin cancer.
7. DERM-ASSESS III
Study Title: DermaSensor Use in the Assessment of Skin Lesions Suggestive of Melanoma III (DERM-ASSESS III)
Study Design: International, multicenter, prospective, blinded clinical study conducted at ten dermatology sites, including eight sites in the United States and two in Australia.
Study Population: 311 subjects with a total of 440 biopsied skin lesions, including 88 high-risk melanocytic lesions.
Study Objective: The study evaluated device performance in classifying biopsied lesions suggestive of melanoma.
Results: Device sensitivity for melanoma was 96% and 91% for high-risk melanocytic lesions (i.e. melanoma, highly atypical nevi). Device specificity was 33%.
Interpretation: The study demonstrated high device sensitivity for melanoma as well as highly atypical nevi. Overall device specificity was 33%, thus DermaSensor could have effectively ruled out 33% of benign lesions biopsied by dermatologists while detecting 96% of melanomas.
8. DermaSensor Repeatability and Reproducibility Study
Study Title: DermaSensor Repeatability and Reproducibility Study (DERM R&R)
Study Design: International, multicenter, prospective, blinded clinical study conducted at two sites, one in the United States and one in Australia.
Study Population: 52 subjects with 64 skin lesions, including 23 malignant lesions.
Study Objective: The study evaluated device repeatability and reproducibility for all lesions and separately for malignant lesions.
Results: Overall repeatability agreement was 84.2%. Overall reproducibility agreement was 87.8%. For malignant lesions only, both repeatability and reproducibility agreement were 97.1%.
Interpretation: The study demonstrated a high level of consistency in device results, particularly for malignant lesions.
9. DermaSensor Pigmented Utility Study
Study Title: A Multi-Reader, Multi-Case Study for Evaluating the Impact of the DermaSensor Device on Primary Care Physicians’ Assessment of Pigmented Lesions
Study Design: U.S.-based, prospective, multi-reader, multi-case study.
Study Population: 77 board-certified PCP readers completed a total of 10,472 assessments.
Study Objective: The study evaluated the performance of PCPs in assessing pigmented lesions when aided by the DermaSensor device result.
Results: Aided PCP sensitivity increased by 6% for all skin cancers. Aided PCP sensitivity for melanoma increased by 7%. Aided PCP accuracy was found to have a non-significant increase, and aided specificity was found to have a non-significant decrease.
Interpretation: The study demonstrated that use of the DermaSensor device results increased PCP sensitivity, including sensitivity for melanoma.
10. DERM-ASSESS III Reader Study
Study Title: A Multi-Reader, Multi-Case Study for Evaluating the Impact of the DermaSensor Device on Primary Care Physicians’ Assessment of Lesions Suggestive of Melanoma
Study Design: U.S.-based, prospective, multi-reader, multi-case study.
Study Population: 118 board-certified PCP readers completed a total of 11,800 assessments.
Study Objective: The study evaluated the performance of PCPs in assessing lesions suggestive of melanoma when aided by the DermaSensor device result.
Results: Aided PCP AUROC increased significantly by 4 percentage points for all skin cancers. Aided PCP sensitivity for all skin cancers increased significantly by 8%. Aided PCP sensitivity for melanoma increased significantly by 9%. Aided PCP specificity decreased non-significantly by 6%.
Interpretation: The study demonstrated improved PCP sensitivity and accuracy when the DermaSensor device result was available, including a significant increase in sensitivity for melanoma.
Overall Summary of the Clinical Evidence
The ten clinical studies described in the DermaSensor IFU evaluate different aspects of the safety, clinical performance, and clinical utility of the DermaSensor device. The clinical effectiveness studies consistently demonstrate high device sensitivity. In studies involving lesions considered by physicians to be suggestive of skin cancer and biopsied accordingly, device sensitivity was always in the range of 90% to 97%.
The DERM-SUCCESS Pivotal Validation Study is particularly important because it included 1,005 subjects and 1,579 pathology-verified lesions from primary care. In this study, DermaSensor achieved a sensitivity of 95.5%. Similarly, the DERM-ASSESS III melanoma validation study demonstrated high sensitivity of 95.5% for melanoma. Taken together, these two studies demonstrate that the device sensitivity is very high and comparable to histopathology, with specificity for physician-biopsied lesions that would enable physicians to avoid 21-33% of unnecessary biopsies.
The utility reader studies address a different and clinically important question: whether physician performance improves when the DermaSensor device result is made available in addition to the physician’s clinical assessment.
Across these studies, availability of the DermaSensor device result consistently increased PCP sensitivity. Depending on the study, sensitivity improved by approximately 6-9%. Meaningful increases in melanoma detection and AUROC were also typically reported, as well as non-significant decreases in specificity.
This reflects the role of DermaSensor as an adjunctive tool designed to support physicians in the assessment of lesions suggestive of skin cancer and to help identify lesions that warrant further evaluation. The DERM Repeatability and Reproducibility Study additionally demonstrated a high level of consistency in device results, particularly for malignant lesions.
Taken together, the clinical evidence supports three principal conclusions:
- High sensitivity for skin cancer detection, including melanoma, basal cell carcinoma, and squamous cell carcinoma.
- Improved sensitivity of Primary Care Physicians when aided by the DermaSensor device result.
- High repeatability and reproducibility of device results particularly for malignant lesions.