Key Points
Question
Does the use of 3-dimensional (3D) total-body photography improve early detection of melanoma and other skin cancers in individuals at high risk of melanoma?
Findings
This randomized clinical trial including 314 patients at high risk of melanoma found that adding 3D total-body photography and sequential digital dermoscopy (via telehealth) to usual clinical surveillance increased the overall number of excisions. However, it did not change the average number of melanomas detected per person in the intervention group compared with the control group.
Meaning
These findings indicate that careful implementation is required to offset increased biopsies of benign lesions in patients at high risk, and further trials are needed in which 3D total-body photography is integrated with teledermatology services or compared with usual care instead of being offered as an add-on service.
Abstract
Importance
Three-dimensional (3D) total-body photography (TBP) can support clinicians in monitoring and identifying changes to skin lesions in patients at high risk of melanoma.
Objective
To assess clinical outcomes between patients at high risk of melanoma receiving usual clinical care compared with those receiving usual care plus 3D TBP and sequential digital dermoscopy imaging (SDDI) every 6 months via teledermatology.
Design, Setting, and Participants
This randomized clinical trial was conducted at a research hospital in Brisbane, Australia, from April 2018 to October 2021, with adult patients (≥18 years) at high risk of developing a primary or subsequent melanoma. Data analysis was conducted from March 2022 to June 2024.
Intervention
Usual care plus 3D-TBP in person and SDDI via teledermatology at baseline, 6, 12, 18, and 24 months. The control group continued usual care and completed online surveys every 6 months.
Main Outcome Measures
Number and rates of excisions and/or biopsies of lesions suggestive of melanoma, and results of histopathologic testing.
Results
The analysis included 314 participants (mean [SD] age, 51.6 [12.8] years; 194 females [62%]) who completed all of the study procedures (158 in the intervention and 156 in the control). In all, 1527 excisions (905 intervention and 622 in the control) were performed among 226 participants (122 intervention and 104 controls), with 67 (4%) histopathologically confirmed as melanoma and 402 (26%) as keratinocyte cancer (KC). The mean (SD) number of lesions of any type excised per person was significantly higher in the intervention (5.73 [6.77]; 95% CI, 4.66-6.79) compared to the control group (3.99 [5.72]; 95% CI, 3.08-4.89; P = .02). Fewer melanomas were detected among the intervention group compared with the control (24 [35%] vs 43 [64%], respectively), and therefore, a lower incidence rate: 2.03 (95% CI, 1.30-3.02) vs 3.62 (95% CI, 2.62-4.88), respectively. After 1 year of follow-up, the intervention had a lower, but not statistically significant, rate of melanoma per person: 0.08 (95% CI, 0.03-0.13) compared with 0.16 (95% CI, 0.08-0.25) in the control; an average of 0.86 (95% CI, 0.55-1.16) vs 0.42 (95% CI, 0.24-0.59) KCs per person; and 2.01 (95% CI, 1.50-2.51) vs 1.39 (95% CI, 0.98-1.82) excisions or biopsies per person, respectively.
Conclusions and Relevance
The results of this randomized clinical trial indicate that the addition of 3D-TPB and SDDI to usual care in a teledermatology setting without AI (artificial intelligence) increased the number and rate of skin excisions and biopsies performed. Further studies are required to compare teledermatology to usual care rather than adding it, and to study whether the use of AI can improve the teledermatology outcomes. Larger studies in multiple settings with a greater number of teledermatologists are needed. This study shows that conducting clinical trials in this setting is feasible.
Trial Registration
anzctr.org.au Identifier: ACTRN12618000267257
This randomized clinical trial compares usual care to usual care plus 3-dimensional total-body photography and sequential digital dermoscopy imaging for early detection of skin cancer in patients at high risk of melanoma.
Introduction
Melanoma represents a substantial and growing health issue. In 2020, an estimated 324 635 new cases of melanoma were diagnosed, and 57 043 melanoma-associated deaths occurred globally.1 Mortality from melanoma increases steeply with increasing tumor thickness at diagnosis, hence early diagnosis is crucial.2 Although potentially beneficial,3 melanoma population screening is not currently recommended because of the lack of clinical evidence for reducing mortality.4 The current recommendation is for self-monitoring and seeking medical advice if changes in skin lesions are noticed; however, few patients report having the confidence to do this well.5,6 For those at high risk in Australia, clinical practice guidelines recommend that individuals undergo a full-skin examination supported by dermoscopy every 6 months, and when possible, total-body photography (TBP) imaging.7 The assessment is conducted at 1 time point and relies on patient and/or clinician recall in the absence of an objective photographic record of changes to lesions. This approach may lead to high rates of excisions or biopsies, as well as the potential for overdiagnosis and overtreatment.8,9
Research suggests that skin surveillance for individuals at high risk can be effective for early detection of melanoma10,11,12; however, most of the evidence to date is from observational studies, which may be prone to bias. TBP can be either 2-dimensional (2D) or 3-dimensional (3D). Sequential 3D TBP allows for fast acquisition of high-resolution macroscopic images of the entire skin surface, which enables clinicians to monitor changes to a patient’s skin.13,14 Dermoscopic images can be linked to the 3D TBP images, and integrated software allows lesions to be sorted by characteristics such as size, color, and change. The 3D TBP images also facilitate assessment via teledermatology.15
The objective of this randomized clinical trial was to compare clinical outcomes (number of skin excisions or biopsies and histopathologic findings) between patients receiving usual clinical care alone compared to usual care plus 3D TBP and sequential digital dermoscopy imaging (SDDI) every 6 months during a 2-year period.
Methods
This study was reviewed and approved by the institutional review board of the Metro South Health Human Research Ethics Committee (HREC/17/QPAH/816). Informed consent was obtained in writing from each participant. The study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. The trial protocol has been published16 and is available in Supplement 1.
Study Design and Participants
This was a 2-arm, single-center, parallel randomized clinical trial conducted at the Clinical Research Facility, Princess Alexandra Hospital (Brisbane, Australia). Participants were enrolled from April 2018 through October 2019 from a university registry of research volunteers and referrals from medical practitioners. Participants were followed up for 2 years, with the last patient visit conducted in October 2021.
Eligible participants were 18 years or older and at high risk of developing melanoma. High risk was defined as having 1 or more of the following: at least 1 melanoma (including in situ) diagnosed before age 40 years; 2 or more melanomas (including in situ) diagnosed before age 65 years; and a strong family history—2 or more first-degree relatives diagnosed with melanoma, a known pathogenic gene variation, and/or a diagnosis of dysplastic nevus syndrome.
All demographic variables, including ancestry, were self-reported in a questionnaire administered by the attending clinician. Other phenotypic characteristics were reported at the time of the visit by the attending clinician. Skin tone was described by clinicians using a 3-point scale (ie, fair, medium, olive).
Randomization
Of 332 individuals assessed, 17 were ineligible and 315 participants were randomly assigned to the intervention or the control group with a 50:50 allocation ratio (159 to the intervention and 156 to the control), using the randomization package embedded within REDCap (Vanderbilt University). Trial staff conducting participant visits were blinded to the allocation sequence. Participants and research staff were aware of group allocation (unblinded). After randomization, 1 participant did not receive the intervention (Figure 1).
Figure 1. Participant Flow Diagram.
Trial Procedures
Intervention Group
Participants randomized to the intervention group first received 3D TBP-SDDI using the VECTRA WB360 (serial number WB00009; Canfield Scientific), and then underwent a clinical skin examination conducted by 1 of 6 junior clinicians every 6 months for 2 years (5 visits total). The 3D TBP device consists of 92 cameras that simultaneously capture images of the entire surface of the patient’s skin. A digital 3D avatar of the patient is constructed from these 92 images.
Images of individual skin nevi (>5 mm) were captured using an attached dermoscopic camera, and lesions were then linked to a digital 3D avatar. A senior teledermatologist (H.P.S.) reviewed images of concerning lesions identified by the junior clinician every 2 weeks in an asynchronous teledermatology model. Any lesion deemed by the senior dermatologist (H.P.S.) to be suggestive of melanoma or other skin cancer was referred for management to the participant’s usual physician, who would determine the treatment course. Participants in the intervention group completed a baseline survey including a melanoma surveillance questionnaire that was repeated at each follow-up visit.
Control Group
Participants in the control group also completed the melanoma surveillance questionnaire at baseline, and were contacted via email or by postal service (mail) every 6 months for 2 years to complete follow-up questionnaires in line with the intervention group visits. All control group participants were offered 3D TBP imaging after completing the last study questionnaire (24 months after study initiation).
Continuation of Usual Care
Participants in both the control and intervention groups were informed by trial staff at baseline that they were considered to be at high risk for melanoma. Therefore, all participants were asked to continue their usual care including regular skin checks by their family physician/general practitioner or dermatologist.
Primary Outcome
The primary outcome was the number and rates of excisions and biopsies and histopathologic findings. With participants’ consent, histopathologic diagnoses of all excised lesions were collected from participants’ physicians or directly from the pathology laboratories. Diagnosis (subtype, stage, and Clark level for melanomas) were extracted from the pathology reports. Community pathologists were not informed of the study. Health economic outcomes based on medical billing data are reported in a companion article.17
Sample Size
To allow for participant withdrawals of approximately 7% (based on our experience with a previous study11), we aimed to recruit 330 participants. This number would provide adequate power to detect a 50% reduction in excision rates in the intervention compared to control group.11,16
Statistical Analysis
The statistical analysis plan is available in Supplement 1. All analyses were performed using an intention-to-treat approach, reporting the number of histopathologically confirmed excisions (total excisions and lesion type) and affected participants. The mean number of excisions per person in the intervention and control groups were compared using t tests. Rates were calculated using recurrent survival analysis, with incidence rates (average excisions/biopsies per 10 000 person-days)18 compared using mid P exact test.19,20 Excision rate ratios (ratio of incidence rate between groups) were reported. Average excisions per participant at 1 year were estimated from the mean cumulative function,18 with a 2-sample pseudo-score test for comparison. The risk of subsequent excisions was estimated using the hazard ratio (HR) from a shared frailty γ model.21 A per-protocol analysis assessed whether missed study visits affected the detection of concerning lesions. Benign to malignant ratios and number needed to excise were reported per study group. All tests conducted were 2-sided and performed using R, version 4.3.2 (The R Foundation for Statistical Computing; reda22 and frailtypack23 packages), and OpenEpi.20 P < .05 was considered statistically significant. Data analyses were performed from March 2022 to June 2024.
Results
Demographic Characteristics
A final sample of 314 participants were followed up for 2 years (158 intervention, 156 control). Baseline characteristics of the 2 groups were similar (mean [SD] age, 51.6 [12.8] years; 194 females [62%] and 120 males [38%]), with innate skin tone described by the attending clinician as fair in 269 patients (86%) and medium in 44 (14%) (Table 1). All participants had either a previous diagnosis of melanoma (302 participants [96%]) or multiple, large, atypical nevi (12 [4%]); and 192 (61%) had a previous keratinocyte cancer (KC) diagnosis. More than three-quarters (80%) of the study population was covered by private health insurance (n = 250).
Table 1. Demographic and Clinical Characteristics of Participants, at Baseline and During Study Period.
| Characteristic | Group, No. (%) | Total (N = 314) | P valuea | |
|---|---|---|---|---|
| Intervention (n = 158) | Control (n = 156) | |||
| Demographic, self-reported at baseline | ||||
| Age, mean (SD), y | 50.9 (13.7) | 52.3 (11.8) | 51.6 (12.8) | .35 |
| Female | 96 (61) | 98 (63) | 194 (62) | .80 |
| Male | 62 (39) | 58 (37) | 120 (38) | |
| Ancestry | ||||
| British or Irish | 145 (92) | 134 (86) | 279 (89) | .38 |
| Other European | 10 (6) | 16 (10) | 26 (8) | |
| Not European | 2 (1) | 3 (2) | 5 (2) | |
| NR | 1 (1) | 4 (2) | 5 (1) | |
| Innate skin tone (ventral upper arm)b | ||||
| Fair | 134 (85) | 135 (87) | 269 (86) | .58 |
| Medium | 23 (15) | 21 (14) | 44 (14) | |
| NR | 1 (1) | 0 | 1 (0.3) | |
| Facultative skin tone (dorsal forearm)b | ||||
| Fair | 61 (39) | 68 (44) | 129 (41) | .59 |
| Medium | 95 (60) | 87 (56) | 182 (58) | |
| Olive | 2 (1) | 1 (1) | 3 (1) | |
| Level of education | ||||
| Primary school | 5 (3) | 6 (4) | 11 (4) | .92 |
| High school degree/certificate | 80 (51) | 76 (49) | 156 (50) | |
| University degree | 72 (46) | 72 (46) | 144 (46) | |
| NR | 1 (1) | 2 (1) | 3 (1) | |
| Employment | ||||
| Employed | 105 (66) | 99 (64) | 204 (65) | .82 |
| Self-employed | 11 (7) | 15 (10) | 26 (8) | |
| Home duties/retired/unemployed | 35 (22) | 36 (23) | 71 (23) | |
| Other | 6 (4) | 4 (3) | 10 (3) | |
| NR | 1 (1) | 2 (1) | 3 (1) | |
| Relationship status | ||||
| De facto/living with partner/married | 130 (82) | 122 (78) | 252 (80) | .47 |
| Single (widowed/divorced/separated/never married) | 27 (17) | 31 (20) | 58 (18) | |
| NR | 1 (1) | 3 (2) | 4 (1) | |
| Private health insurance | ||||
| Yes | 129 (82) | 121 (78) | 250 (80) | .47 |
| No | 28 (18) | 32 (21) | 60 (19) | |
| NR | 1 (1) | 3 (2) | 4 (1) | |
| Clinical, survey responses at baseline | ||||
| “Has a doctor ever diagnosed you with melanoma?” | ||||
| Yes | 150 (95) | 152 (97) | 302 (96) | .06 |
| No | 8 (5) | 2 (1) | 10 (3) | |
| NR | 0 | 2 (1) | 2 (1) | |
| “Has a doctor ever diagnosed you with nonmelanoma skin cancers?” | ||||
| Yes | 95 (60) | 97 (62) | 192 (61) | .76 |
| No | 62 (39) | 57 (37) | 119 (38) | |
| NR | 1 (1) | 2 (1) | 3 (1) | |
| “How often do you get skin checks?” | ||||
| More often than every 3 mo | 5 (3) | 6 (4) | 11 (4) | .06 |
| Every 3-5 mo | 61 (39) | 50 (32) | 111 (35) | |
| Every 6-11 mo | 65 (41) | 53 (34) | 118 (38) | |
| Every 12 mo | 20 (13) | 28 (18) | 48 (15) | |
| Every 24 mo | 0 | 7 (5) | 7 (2) | |
| Less often than every 24 mo | 1 (1) | 2 (1) | 3 (1) | |
| NR | 6 (4) | 10 (6) | 16 (5) | |
| “Do you have a first-degree relative with a history of melanoma?” | ||||
| Yes | 104 (66) | 88 (56) | 192 (61) | .07 |
| No | 40 (25) | 59 (38) | 99 (32) | |
| “Don’t know” | 13 (8) | 7 (5) | 20 (6) | |
| NR | 1 (1) | 2 (1) | 3 (1) | |
| Clinical, during study period | ||||
| Melanomas developed | ||||
| 0 | 139 (88) | 127 (81) | 266 (85) | .14 |
| 1 | 16 (10) | 20 (13) | 36 (12) | |
| 2 | 1 (1) | 6 (4) | 7 (2) | |
| 3 | 2 (1) | 1 (1) | 3 (1) | |
| 4 | 0 | 2 (1) | 2 (1) | |
| Nevi on 3D TBD (automated count), mean (SD) | 140.9 (120.1)c | 155.2 (147.5)d | 147.4 (133.1) | .38 |
Abbreviations: 3D TBP, 3-dimensional total-body photography; NR, not reported; SDDI, sequential digital dermoscopy imaging.
χ2/Fisher exact test for comparing count data between 2 groups; 2-sample t test for comparing means.
Clinician described using a 3-point scale (ie, fair, medium, olive).
For comparability with the control group nevus count, counts were obtained from the last available 3D TBP image of each participant.
Based on 129 participants who opted to undergo 3D TBP-SDDI offered to them after study completion.
Number of Excisions and Histopathologic Diagnoses
Table 2 presents the number of excisions and excision rates overall for both melanoma and KC. eTable 1 in Supplement 2 presents additional details on KC subtypes and benign lesions.
Table 2. Number and Rates of Lesions Excised, Total and Per Category, by Study Group.
| Category | No. (95% CI)a | P value | |
|---|---|---|---|
| Intervention group | Control group | ||
| Participants, No. | 158 | 156 | NA |
| Total excisions (any reason) | |||
| Total person-days | 118 246 | 118 691 | NA |
| Total excisions, No. | 905 | 622 | NA |
| Participants with excisions, No. | 122 | 104 | NA |
| Mean excisions per person | 5.73 (4.66-6.79) | 3.99 (3.08-4.89) | .02 |
| Incidence rate | 76.3 (71.4-81.4) | 52.6 (48.6-56.9) | <.001 |
| Rate ratio | 1.46 (1.32-1.62) | 1 [Reference] | |
| Average rate per person at 1 y | 2.94 (2.27-3.62) | 1.97 (1.44-2.50) | .007b |
| HR for risk of subsequent excisions | 1.60 (1.23-2.06)c | 1 [Reference] | <.001 |
| Malignant lesions, No. (% of total excisions)d | 280 (31) | 189 (30) | NA |
| Benign to malignant ratio | 2.2:1.0 (625:280) | 2.3:1.0 (433:189) | NA |
| No. needed to excise | 3.2 (3.1-3.3) | 3.3 (2.8-3.9) | NA |
| All melanomas | |||
| No. (% of total excisions) | 24 (3) | 43 (7) | NA |
| Participants with melanoma, No. (%) | 19 (12) | 29 (19) | NA |
| Mean melanomas per person | 0.15 (0.08-0.23) | 0.28 (0.17-0.38) | .06 |
| Incidence rate | 2.03 (1.30-3.02) | 3.62 (2.62-4.88) | .02 |
| Rate ratio | 0.56 (0.34-0.92) | 1 [Reference] | |
| Average rate per person at 1 y | 0.08 (0.03-0.13) | 0.16 (0.08-0.25) | .08b |
| HR for risk of subsequent melanomas | 0.63 (0.34-1.14)c | 1 [Reference] | .13 |
| Melanoma in situ | |||
| No. (% of total excisions) | 20 (2) | 34 (5) | NA |
| Participants with melanoma in situ, No. (%) | 15 (9) | 25 (16) | NA |
| Mean melanomas in situ per person | 0.13 (0.06-0.20) | 0.22 (0.13-0.31) | .12 |
| Incidence rate | 1.69 (1.03-2.61) | 2.87 (1.98-4.00) | .06 |
| Rate ratio | 0.59 (0.34-1.03) | 1 [Reference] | |
| Average rate per person at 1 y | 0.06 (0.01-0.11) | 0.14 (0.06-0.21) | .09b |
| HR for risk of subsequent melanoma in situ | 0.60 (0.31-1.17) | 1 [Reference] | .14 |
| Invasive melanoma | |||
| No. (% of total excisions) | 4 (0.4) | 9 (1) | NA |
| Participants with invasive melanomas, No. (%) | 4 (3) | 7 (4) | NA |
| Mean invasive melanomas per person | 0.03 (0-0.05) | 0.06 (0.01-0.10) | .21 |
| Incidence rate | 0.34 (0.09-0.87) | 0.76 (0.35-1.44) | .18 |
| Rate ratio | 0.45 (0.14-1.45) | 1 [Reference] | |
| Average rate per person at 1 y | 0.02 (0.00-0.04) | 0.03 (0.00-0.05) | .52b |
| HR for risk of subsequent invasive melanomas | 0.46 (0.14-1.45) | 1 [Reference] | .18 |
| All keratinocyte cancers | |||
| No. (% of total excisions) | 256 (28) | 146 (23) | NA |
| Participants with keratinocyte cancers, No. (%) | 59 (37) | 44 (28) | NA |
| Mean keratinocyte cancers per person | 1.62 (1.11-2.13) | 0.94 (0.58-1.29) | .03 |
| Incidence rate | 21.65 (19.08-24.47) | 12.30 (10.39-14.47) | <.001 |
| Rate ratio | 1.76 (1.44-2.16) | 1 [Reference] | |
| Average rate per person at 1 y | 0.86 (0.55-1.16) | 0.42 (0.24-0.59) | .01b |
| HR for risk of subsequent keratinocyte cancers | 2.08 (1.33-3.27)c | 1 [Reference] | <.001 |
| Benign lesions excised | |||
| No. (% of total excisions) | 625 (69) | 433 (70) | NA |
| Participants with benign lesions excised, No. (%) | 120 (76) | 92 (59) | NA |
| Mean benign lesion excisions per person | 3.96 (3.16- 4.75) | 2.78 (2.04-3.50) | .03 |
| Incidence rate | 52.86 (48.79-57.17) | 36.48 (33.13-40.08) | <.001 |
| Rate ratio | 1.45 (1.28-1.64) | 1 [Reference] | |
| Average excision rate per person at 1 y | 2.01 (1.50-2.51) | 1.39 (0.98-1.82) | .02b |
| HR for risk of subsequent benign lesion excisions | 1.50 (1.16-1.93)c | 1 [Reference] | <.001 |
Abbreviations: HR, hazard rate; NA, not applicable; NR, not reported.
Any negative lower bound of a CI was reported as 0.
Although the average rate of lesions per person is reported for 1-year follow-up, the P value is from a 2-sample pseudo-score test to test; there was no difference between the average rate of lesions per person between the intervention and the control for the total study duration.
Model adjusted for age and sex, and the P value reported is for the treatment group variable in the model. Models for subtypes of lesions were not adjusted for age and sex for consistency because some groups with a small sample size had convergence problems when age and sex were included in the model.
All histopathologically confirmed excised lesions. Melanomas were classified according to the American Joint Committee on Cancer Staging Manual, eighth edition,24 as either in situ or invasive.
The median (IQR; total person-days) total follow-up time was 2.02 years (1.98-2.11; 118 691) for the intervention group, and 2.05 years (2.00-2.18; 118 246) for the control.
A total of 1527 excisions were conducted in 226 participants (122 intervention; 104 control). The mean (SD) number of lesions of any type excised per person was significantly higher in the intervention (5.73 [6.77]; 95% CI, 4.66-6.79) compared to the control group (3.99 [5.72]; 95% CI, 3.08-4.89; P = .02). When considering only participants with at least 1 lesion excised, the mean (SD) number of excisions per person was not significantly different between the intervention (7.42 [6.84]; 95% CI, 6.19-8.64) and the control group (5.98 [6.10]; 95% CI, 4.79-7.17; P = .10). eTable 2 in Supplement 2 describes health care practitioners who completed each excision.
Of the 1527 total excisions, 67 (4%) were histopathologically confirmed melanoma, diagnosed in 48 participants (7%), 19 in the intervention and 29 in the control group (Table 2). Of the melanomas, 54 (81%) were in situ, and 13 (19%) were invasive, with only 1 with a thickness greater than 0.8 mm. Melanoma in situs included 16 (30%) superficial spreading, 13 (24%) lentiginous, and 3 (6%) lentigo maligna; the remaining 22 (41%) melanomas were not further specified. The 13 invasive melanomas included 7 (54%) superficial spreading and 1 (8%) lentiginous subtype; 5 (38%) were not further specified. Twelve participants (5%) had multiple melanomas (3 intervention; 9 control; Table 1). Regarding body site, melanomas were most frequently found on the back (n = 23 [34%]) and lower leg (n = 17 [25%]). Of the 1527 total excisions, 402 (26%) were KCs histopathologically confirmed in 59 intervention participants and in 44 control participants, and included 215 basal cell carcinomas (14%), 160 intra-epidermal carcinomas (10%), and 27 squamous cell carcinomas (2%). KCs were most frequently found on the head (100 [25%]) and the back (66 [16%]).
Of the 1527 total excisions, 599 (39%) were histopathologically confirmed nevi, equally distributed among the intervention (357 of 905 excisions [38%]) and control group (242 of 622 excisions [39%]). Most of the excised nevi among the intervention group were on the back (n = 145; 41%), and in the control group, on the legs (n = 88; 36%).
Of the 908 total excisions, 147 (16%) in the intervention group were referred through the study, and 4 were melanomas.
Excision Rates Over Time
Total Lesion Excision Rate
The incidence rate of excisions during follow-up in the intervention group (76.3; 95% CI, 71.4-81.4) was approximately 1.45 times (95% CI, 1.32-1.62) the rate of the control group (52.6; 95% CI, 48.6-56.9; P < .001) (Table 2). At the 1-year follow-up, an average of 2.94 (95% CI, 2.27-3.62) and 1.97 (95% CI, 1.44-2.50) excisions per person were experienced in the intervention and control groups, respectively. In the model adjusted for age and sex, the risk of having subsequent excisions was 60% higher among the intervention compared to the control groups (HR, 1.60; 95% CI, 1.23-2.06; P < .001) (Table 2).
Melanoma Excision Rate
The incidence rate of melanomas during follow-up period in the intervention group (2.03; 95% CI, 1.30-3.02) was 0.56 times (95% CI, 0.34-0.92) the rate of the control group (3.62; 95% CI, 2.62-4.88; P = .02). Number of melanomas diagnosed per person at the 1-year follow-up was 0.08 (95% CI, 0.03-0.13) and 0.16 (95% CI, 0.08-0.25) (Table 2; Figure 2) in the intervention and control groups, respectively. In the intervention group, 3 of 158 participants (2%) had more than 1 melanoma during the study period, compared to 9 of 156 participants (6%) in the control group. The risk of having a subsequent melanoma was not significantly different for the intervention group compared to the control group (HR, 0.63; 95% CI, 0.34-1.14; P = .13) (Table 2).
Figure 2. Mean Cumulative Function (MCF) Estimates for the Average Number of Melanomas per Person at Each Time Point.
Shaded areas represent 95% CIs and hatch marks indicate right censoring for each participant on the last date in the study. All participants were at risk at each time point (measured in number of days), hence number at risk was not denoted below the x-axis. The average number of melanomas per person was greater among the control group than among the intervention group.
Keratinocyte Cancer Excision Rate
The incidence rate of KCs during follow-up in the intervention group (21.65; 95% CI, 19.08-24.47) was 1.76 times (95% CI, 1.44-2.16) the rate of the control group (12.30; 95% CI, 10.39-14.47; P < .001). The number of KCs diagnosed per person at the 1-year follow-up was 0.86 (95% CI, 0.55-1.16) and 0.42 (95% CI, 0.24-0.59) in the intervention and control groups, respectively (Table 2). The HR for subsequent KCs was 108% higher in the intervention compared to the control group (HR, 2.08; 95% CI, 1.33-3.27; P < .001).
Benign to Malignant Ratio
The benign to malignant ratio for excised lesions did not differ between the groups. The ratio was 2.2:1.0 (625 to 280) in the intervention group, and 2.3:1.0 (433 to 189) in the control group.
Number Needed to Excise
The number needed to excise was similar, 3.2 in the intervention group (95% CI, 3.1-3.3) and 3.3 in the control group (95% CI, 2.8-3.9).
The per-protocol analysis censoring participants from the intervention who missed at least 1 study visit showed no difference in the average number of excisions per person, except for the average number of nevi excised, between participants attending all study visits compared to the controls (eTable 3 in Supplement 2).
Discussion
Our analysis found that adding 3D TBP-SDDI and teledermatology reviews to usual care compared with usual care alone resulted in an increased number of excisions or biopsies for individuals at high risk of melanoma, contrary to the hypothesis of a decrease. There was a higher overall mean number of excisions and average excisions per-person rate in the intervention group. The most frequently excised lesions were KCs and nevi.
A high number of subsequent in situ melanomas were found. The higher rate of excisions in the intervention group may be associated with the additional surveillance received by that group compared with the control group participants who visited only their usual physician. Clinician attention may be heightened to detect even subtle lesion irregularities detected by surveillance photography. This aligns with evidence that indicated people who underwent skin screening subsequently experienced higher rates of biopsies, and if a melanoma was detected, they were frequently thinner compared to those detected outside of a skin screening examination.25 The overall low number of melanomas detected in the intervention group may be due to a random difference between the 2 groups or heightened sensitivity in which potential melanomas (precursor lesions) are removed before they develop further. Conversely, it is possible that some melanomas were overlooked because TBP was reviewed using a teledermatology model, as opposed to a whole-body skin examination by a consultant dermatologist. However, this seems unlikely given that all participants continued with usual care and the high number of KCs detected. Furthermore, the per-person excision rates were significantly different between the groups only when the risk of having subsequent lesions excised was highly significant, implying a few participants with a high number of excisions may have induced the higher excision rates in the intervention group.
Population screening for melanoma is contentious due to a lack of evidence from randomized clinical trials on mortality reduction.4,26,27,28 Currently, screening is opportunistic, with approximately one-third of adults in Australia receiving a skin check annually.29 Targeted melanoma surveillance, including TBP-SDDI in a high-risk clinic model, has shown promise: in patients at high risk, melanomas detected by TBP are thinner than those detected by traditional methods.30,31 Before TBP can be widely recommended, numerous challenges remain that should be addressed: integration of TBP into clinical workflows, who should be screened, privacy and confidentiality, and cost and regulatory approvals, among others.12,32,33,34,35 There is concern that increased screening and subsequent biopsies combined with shifts in pathological thresholds are producing overdiagnosis of melanoma.36,37,38 Overdiagnosis has implications for overtreatment, patient harm, and increased health care costs. Another concern with screening is the potential psychosocial impact and the “worried well,” ie, those who are relatively healthy and at low risk but consider themselves likely to be affected, and therefore, overutilize screening services.39 A positive aspect of screening in Australia is that the general population reports high acceptance of 3D TBP-SDDI.40,41,42
Although artificial intelligence (AI) was not used in this study, our data are contributing to current development efforts.43,44,45 Several studies of AI in dermatology report high-level performance indicating its potential clinical use.46,47,48 Currently, the Australian College of Dermatologists Position Statement49 recommends that only AI models that have regulatory approval from the Therapeutic Goods Administration post-2021 be used in clinical practice. In the future, AI tools embedded in 3D TBP-SDDI could facilitate patient risk stratification for targeted screening, triage lesions, and/or provide a second opinion for dermatologists.50 The highest perceived benefit of AI reported by dermatologists is more efficient triaging; however, future studies will need to test the effect on excision rates.51
Limitations
We note the following limitations. Junior clinicians triaged concerning lesions for the teledermatologist to review; therefore, the lesion subset flagged was dependent on the junior clinicians’ expertise. This study was conducted during the height of the COVID-19 pandemic and its public health restrictions, which may have affected participants’ screening habits and attendance at study visits.52 We did not ask participants if they received any imaging outside the study. Given that this was a research study and not intended to replace usual clinical care, participants in both groups were advised to continue to see their usual physician or dermatologist for routine skin checks, which likely resulted in the participants being more diligent with their skin screenings. The participants’ usual physicians were aware that these patients were participating in this study, and that the referring specialist was a senior dermatologist, which potentially influenced their decisions to excise lesions (Hawthorne effect53). The usual care physicians making the clinical management decisions about the excisions did not have access to the 3D TBP images, but rather the macro (and dermoscopic if taken) images via a referral letter. A major limitation is that this reduced the benefit of the collection of longitudinal images because these clinicians were not able to see whether the lesions had changed over time. In the future, we recommend that the 3D TBP be provided to the treating physician to assist in clinical decision-making to align with the intended clinical workflow using this technology. Furthermore, excised lesions were reviewed by different communities of pathologists, and this may have affected the threshold by which a lesion is considered to be malignant. Moreover, our data lacked diversity given the predominantly fair-skinned population. Likewise, most participants had private insurance, and therefore, these findings may have limited generalizability to other populations. Data show that biopsies, melanoma incidence, and associated health system costs are higher in Queensland compared with other Australian states.9,54 Finally, it is possible that randomization of participants may not have resulted in similar distribution of unmeasured preexisting characteristics across the 2 groups as intended.
Conclusions
This randomized clinical trial found that adding sequential 3D TBP-SDDI skin examinations to a patient’s usual care in a teledermatology setting led to an increase in excisions among patients at high risk for melanoma. Moreover, there was no significant change in the number of melanomas diagnosed per person among the intervention participants who received 3D TBP-SDDI skin examinations every 6 months for 2 years and those who continued usual care. Further studies are required to compare teledermatology to clinical services. This study did not use AI; however, future studies would benefit from exploring its role in telehealth models.
Trial Protocol and Statistical Analysis Plan (after amendments to the plan in the protocol)
eTable 1. Number and Rates of Lesions Excised per Category, Continued From Table 2
eTable 2. Health Care Practitioners Who Completed Excisions
eTable 3. Number and Rates of Lesions Excised, per Protocol Analysis
Data Sharing Statement
References
- 1.Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249. doi: 10.3322/caac.21660 [DOI] [PubMed] [Google Scholar]
- 2.Gershenwald JE, Scolyer RA, Hess KR, et al. Melanoma staging: evidence-based changes in the American Joint Committee on Cancer eighth edition cancer staging manual. CA Cancer J Clin. 2017;67(6):472-492. doi: 10.3322/caac.21409 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tripp MK, Watson M, Balk SJ, Swetter SM, Gershenwald JE. State of the science on prevention and screening to reduce melanoma incidence and mortality: the time is now. CA Cancer J Clin. 2016;66(6):460-480. doi: 10.3322/caac.21352 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mangione CM, Barry MJ, Nicholson WK, et al. ; US Preventive Services Task Force . Screening for skin cancer: US Preventive Services Task Force Recommendation Statement. JAMA. 2023;329(15):1290-1295. doi: 10.1001/jama.2023.4342 [DOI] [PubMed] [Google Scholar]
- 5.Dieng M, Smit AK, Hersch J, et al. Patients’ views about skin self-examination after treatment for localized melanoma. JAMA Dermatol. 2019;155(8):914-921. doi: 10.1001/jamadermatol.2019.0434 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ackermann DM, Bracken K, Janda M, et al. Strategies to improve adherence to skin self-examination and other self-management practices in people at high risk of melanoma: a scoping review of randomized clinical trials. JAMA Dermatol. 2023;159(4):432-440. doi: 10.1001/jamadermatol.2022.6478 [DOI] [PubMed] [Google Scholar]
- 7.Cancer Council Australia . Clinical Guidelines. Accessed October 30, 2024. https://wiki.cancer.org.au/australia/Clinical_question:What_interventions_reduce_risk_of_death_from_melanoma_in_high_risk_people%3F
- 8.Gordon LG, Elliott TM, Olsen CM, Pandeya N, Whiteman DC. Multiplicity of skin cancers in Queensland and their cost burden to government and patients. Aust N Z J Public Health. 2018;42(1):86-91. doi: 10.1111/1753-6405.12738 [DOI] [PubMed] [Google Scholar]
- 9.Gordon LG, Leung W, Johns R, et al. Estimated healthcare costs of melanoma and keratinocyte skin cancers in Australia and Aotearoa New Zealand in 2021. Int J Environ Res Public Health. 2022;19(6):3178. doi: 10.3390/ijerph19063178 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Watts CG, Cust AE, Menzies SW, Coates E, Mann GJ, Morton RL. Specialized surveillance for individuals at high risk for melanoma: a cost analysis of a high-risk clinic. JAMA Dermatol. 2015;151(2):178-186. doi: 10.1001/jamadermatol.2014.1952 [DOI] [PubMed] [Google Scholar]
- 11.Watts CG, Cust AE, Menzies SW, Mann GJ, Morton RL. Cost-effectiveness of skin surveillance through a specialized clinic for patients at high risk of melanoma. J Clin Oncol. 2017;35(1):63-71. doi: 10.1200/JCO.2016.68.4308 [DOI] [PubMed] [Google Scholar]
- 12.Hornung A, Steeb T, Wessely A, et al. The value of total body photography for the early detection of melanoma: a systematic review. Int J Environ Res Public Health. 2021;18(4):1726. doi: 10.3390/ijerph18041726 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rayner JE, Laino AM, Nufer KL, et al. Clinical perspective of 3D total body photography for early detection and screening of melanoma. Front Med (Lausanne). 2018;5:152. doi: 10.3389/fmed.2018.00152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bhave P, Wong J, McInerney-Leo A, et al. Management of cutaneous melanoma in Australia: a narrative review. Med J Aust. 2023;218(9):426-431. doi: 10.5694/mja2.51910 [DOI] [PubMed] [Google Scholar]
- 15.Jones LK, Oakley A. Store-and-forward teledermatology for assessing skin cancer in 2023: literature review. JMIR Dermatol. 2023;6:e43395. doi: 10.2196/43395 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Primiero CA, McInerney-Leo AM, Betz-Stablein B, et al. Evaluation of the efficacy of 3D total-body photography with sequential digital dermoscopy in a high-risk melanoma cohort: protocol for a randomised controlled trial. BMJ Open. 2019;9(11):e032969. doi: 10.1136/bmjopen-2019-032969 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lindsay D, Soyer HP, Janda M, et al. Cost-effectiveness analysis of 3D total-body photography for people at high risk of melanoma. JAMA Dermatol. Published online March 26, 2025. doi: 10.1001/jamadermatol.2025.0219 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kuramoto L, Sobolev BG, Donaldson MG. On reporting results from randomized controlled trials with recurrent events. BMC Med Res Methodol. 2008;8(1):35. doi: 10.1186/1471-2288-8-35 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Martin DO, Austin H. Exact estimates for a rate ratio. Epidemiology. 1996;7(1):29-33. doi: 10.1097/00001648-199601000-00006 [DOI] [PubMed] [Google Scholar]
- 20.Dean AGSK, Soe MM. OpenEpi: Open Source Epidemiologic Statistics for Public Health. 3.01 ed. 2013. [Google Scholar]
- 21.Box-Steffensmeier JM, De Boef S. Repeated events survival models: the conditional frailty model. Stat Med. 2006;25(20):3518-3533. doi: 10.1002/sim.2434 [DOI] [PubMed] [Google Scholar]
- 22.Wang WFH, Yan J. reda: Recurrent Event Data Analysis. 2022. Accessed February 19, 2024. https://cran.r-project.org/web/packages/reda/reda.pdf
- 23.Virginie Rondeau YM, Gonzalez JR. Frailty pack: an R package for the analysis of correlated survival data with frailty models using penalized likelihood estimation or parametrical estimation. J Stat Softw. 2012;47(4):1-28. [Google Scholar]
- 24.Keung EZ, Gershenwald JE. The eighth edition American Joint Committee on Cancer (AJCC) melanoma staging system: implications for melanoma treatment and care. Expert Rev Anticancer Ther. 2018;18(8):775-784. doi: 10.1080/14737140.2018.1489246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Whiteman DC, Olsen CM, MacGregor S, et al. ; QSkin Study . The effect of screening on melanoma incidence and biopsy rates. Br J Dermatol. 2022;187(4):515-522. doi: 10.1111/bjd.21649 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lew JB, Feletto E, Wade S, et al. Benefits, harms and cost-effectiveness of cancer screening in Australia: an overview of modelling estimates. Public Health Res Pract. 2019;29(2):29121913. doi: 10.17061/phrp2921913 [DOI] [PubMed] [Google Scholar]
- 27.Janda M, Cust AE, Neale RE, et al. Early detection of melanoma: a consensus report from the Australian Skin and Skin Cancer Research Centre Melanoma Screening Summit. Aust N Z J Public Health. 2020;44(2):111-115. doi: 10.1111/1753-6405.12972 [DOI] [PubMed] [Google Scholar]
- 28.Janda M, Olsen CM, Mar J, Cust AE. Early detection of skin cancer in Australia - current approaches and new opportunities. Public Health Res Pract. 2022;32(1):3212204. doi: 10.17061/phrp3212204 [DOI] [PubMed] [Google Scholar]
- 29.Reyes-Marcelino G, Tabbakh T, Espinoza D, et al. Prevalence of skin examination behaviours among Australians over time. Cancer Epidemiol. 2021;70:101874. doi: 10.1016/j.canep.2020.101874 [DOI] [PubMed] [Google Scholar]
- 30.Rademaker M, Oakley A. Digital monitoring by whole body photography and sequential digital dermoscopy detects thinner melanomas. J Prim Health Care. 2010;2(4):268-272. doi: 10.1071/HC10268 [DOI] [PubMed] [Google Scholar]
- 31.Watts CG, Wortley S, Norris S, et al. A national budget impact analysis of a specialised surveillance programme for individuals at very high risk of melanoma in Australia. Appl Health Econ Health Policy. 2018;16(2):235-242. doi: 10.1007/s40258-017-0368-0 [DOI] [PubMed] [Google Scholar]
- 32.Horsham C, Janda M, Kerr M, Soyer HP, Caffery LJ. Consumer perceptions on privacy and confidentiality in dermatology for 3D total-body imaging. Australas J Dermatol. 2023;64(1):118-121. doi: 10.1111/ajd.13952 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Nguyen J, Ting S, Paul E, et al. Diagnostic tools used for melanoma: a survey of Australian general practitioners and dermatologists. Australas J Dermatol. 2021;62(3):300-309. doi: 10.1111/ajd.13595 [DOI] [PubMed] [Google Scholar]
- 34.Mahumud RA, Janda M, Soyer HP, Fernández-Peñas P, Mar VJ, Morton RL. Assessing the value of precision medicine health technologies to detect and manage melanoma. Med J Aust. 2022;217(6):275-278. doi: 10.5694/mja2.51696 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lee KJ, Betz-Stablein B, Stark MS, et al. The future of precision prevention for advanced melanoma. Front Med (Lausanne). 2022;8:818096. doi: 10.3389/fmed.2021.818096 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Bjørch MF, Gram EG, Brodersen JB. Overdiagnosis in malignant melanoma: a scoping review. BMJ Evid Based Med. 2023. [DOI] [PubMed] [Google Scholar]
- 37.Matsumoto M, Wack S, Weinstock MA, et al. Five-year outcomes of a melanoma screening initiative in a large health care system. JAMA Dermatol. 2022;158(5):504-512. doi: 10.1001/jamadermatol.2022.0253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Betz-Stablein B, Soyer HP. Overdiagnosis in melanoma screening: is it a real problem? Dermatol Pract Concept. 2023;13(4):e2023247. doi: 10.5826/dpc.1304a247 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Kim A, Chung KC, Keir C, Patrick DL. Patient-reported outcomes associated with cancer screening: a systematic review. BMC Cancer. 2022;22(1):223. doi: 10.1186/s12885-022-09261-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Horsham C, O’Hara M, Sanjida S, et al. The experience of 3D total-body photography to monitor nevi: results from an Australian general population-based cohort study. JMIR Dermatol. 2022;5(2):e37034. doi: 10.2196/37034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Koh U, Janda M, Aitken JF, et al. ‘Mind your moles’ study: protocol of a prospective cohort study of melanocytic naevi. BMJ Open. 2018;8(9):e025857. doi: 10.1136/bmjopen-2018-025857 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Hona TWPT, Horsham C, Silva CV, et al. ; ACEMID group . Consumer views of melanoma early detection using 3D total-body photography: cross-sectional survey. Int J Dermatol. 2023;62(4):524-533. doi: 10.1111/ijd.16578 [DOI] [PubMed] [Google Scholar]
- 43.Melanoma Institute Australia . Risk Predictor Tools. Accessed February 14, 2024. https://www.melanomarisk.org.au/index
- 44.QIMR Berghofer Medical Research Institute . Melanoma Risk Predictor. Accessed February 14, 2024 https://publications.qimrberghofer.edu.au/Custom/QSkinMelanomaRisk
- 45.Koh U, Cust AE, Fernández-Peñas P, et al. ACEMID cohort study: protocol of a prospective cohort study using 3D total body photography for melanoma imaging and diagnosis. BMJ Open. 2023;13(9):e072788. doi: 10.1136/bmjopen-2023-072788 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Zhang S, Wang Y, Zheng Q, Li J, Huang J, Long X. Artificial intelligence in melanoma: a systematic review. J Cosmet Dermatol. 2022;21(11):5993-6004. doi: 10.1111/jocd.15323 [DOI] [PubMed] [Google Scholar]
- 47.Li S, Chu Y, Wang Y, et al. Distinguish the value of the benign nevus and melanomas using machine learning: a meta-analysis and systematic review. Mediators Inflamm. 2022;2022:1734327. doi: 10.1155/2022/1734327 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Tschandl P, Rinner C, Apalla Z, et al. Human-computer collaboration for skin cancer recognition. Nat Med. 2020;26(8):1229-1234. doi: 10.1038/s41591-020-0942-0 [DOI] [PubMed] [Google Scholar]
- 49.Caffery LJ, Janda M, Miller R, et al. Informing a position statement on the use of artificial intelligence in dermatology in Australia. Australas J Dermatol. 2023;64(1):e11-e20. doi: 10.1111/ajd.13946 [DOI] [PubMed] [Google Scholar]
- 50.Primiero CA, Rezze GG, Caffery LJ, et al. A narrative review: opportunities and challenges in artificial intelligence skin image analyses using total body photography. J Invest Dermatol. 2024;144(6):1200-1207. doi: 10.1016/j.jid.2023.11.007 [DOI] [PubMed] [Google Scholar]
- 51.Nelson CA, Pachauri S, Balk R, et al. Dermatologists’ perspectives on artificial intelligence and augmented intelligence: a cross-sectional survey. JAMA Dermatol. 2021;157(7):871-874. doi: 10.1001/jamadermatol.2021.1685 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Canfell K, Chiam K, Nickson C, Mann GB. The complex impact of COVID-19 on cancer outcomes in Australia. Med J Aust. 2023;219(9):402-404. doi: 10.5694/mja2.52125 [DOI] [PubMed] [Google Scholar]
- 53.McCambridge J, Witton J, Elbourne DR. Systematic review of the Hawthorne effect: new concepts are needed to study research participation effects. J Clin Epidemiol. 2014;67(3):267-277. doi: 10.1016/j.jclinepi.2013.08.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Cramb SM, Duncan EW, Aitken JF, et al. Geographical patterns in melanoma incidence across Australia: can thickness differentials reveal the key drivers? Ann Cancer Epidemiol. 2020:4. doi: 10.21037/ace-20-13 32954254 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol and Statistical Analysis Plan (after amendments to the plan in the protocol)
eTable 1. Number and Rates of Lesions Excised per Category, Continued From Table 2
eTable 2. Health Care Practitioners Who Completed Excisions
eTable 3. Number and Rates of Lesions Excised, per Protocol Analysis
Data Sharing Statement
