The patients were men (n
= 81) and women (n
= 120) ranging in age from 7 to 91 (mean age was 60.3 ± 16.3 versus 54.6 ± 15.6, resp.). Ninety one percent (91%) of patients had a skin type of II, and 37% had a history of blistering sunburn. Over a 10-year period, n
= 101, melanomas were initially identified by the dermatologist (DI group), with the rest initially identified by the patient or family members (PI group; n
= 100) (). In three cases, patients were adopted and family history could not be determined, with the remaining subjects (n
= 198) distributed among those with no family history (n
= 121, 61%), a family history of melanoma (n
= 30, 15%), and a family history of other types of skin cancer (n
= 47, 24%). Personal histories were divided among subjects with no history (n
= 89, 44%), and those with a history of melanoma (n
= 18, 9%), atypical nevi (n
= 20, 10%), and other types of skin cancer (n
= 74, 37%). Overall, 158 (78.5%) subjects had melanoma in situ, 26 (13%) a Clark's Level II tumor, 12 (6%) a Clark's Level III, 4 (2%) a Level IV, and 1 (0.5%) a Level V tumor. The average Breslow's thickness for the 43 subjects in which it was measured was 0.73
mm. Twenty-seven subjects had a Breslow's thickness ≥0.75, and 16 subjects <0.75 (). The average Breslow thickness in the PI group was 0.79
mm (SD = 0.73
= 33) and 0.53
mm in the DI group (SD = 0.26
Analysis of age yielded significant differences between the PI (mean age = 54.7
yrs, SD = 12.5) and DI groups (mean age = 60.7
yrs, SD = 16.7), (t
= 2.9, P
< 0.005), (). Diagnostic source was also significantly related to a family history of melanoma, with significantly more melanomas identified by patients when there was a family history of melanoma (DI = 7%, PI = 23%) compared to no family history (DI = 67%, PI
= 55%), or a family history of other skin cancers (DI = 26%, PI = 22%),
= 9.7, P
< 0.01). No such association was found for a personal history of skin cancer (χ2
= 9.5, P
< 0.05, n.s.). No between-group differences were found for gender, skin type, and history of blistering sunburns.
Differences in tumor types, severity, and location between DI and PI groups are provided in . There was a significant relationship between diagnostic source and tumor location with DI tumors more likely to be present on the chest/trunk (DI = 19%, PI = 14%), back (DI = 19%, PI = 14%), and legs (DI = 34%, PI = 21%), while PI tumors were more frequently identified on the face/neck (DI = 9%, PI = 29%) and scalp (DI = 1%, PI = 2%), (χ2 = 15.4, P < 0.01). No between-group differences were found for gender, skin type, and history of blistering sunburns.
Analysis of diagnostic source and tumor severity resulted in significantly less invasive melanomas in the DI group compared to the PI group (χ2 = 15.9, P < 0.0005). Examination of tumor type showed that for DI tumors, 90% were diagnosed as melanoma in situ, 7% were identified as Clark's level II, 2% level III, and 1% level IV. In contrast, PI tumors were 67% in situ, 19% Clark's level II, 10% level III, and 4% level IV.
The post-hoc logistic regression analysis revealed that only one significant variable predicted tumor invasiveness. Specifically, of all the predictor variables (age, family history, personal history, tumor location, and diagnostic source), only diagnostic source (i.e., PI versus DI) was a significant predictor of tumor invasiveness (OR = 0.19; 95% CI, 0.08–0.43; P < 0.0001).