This dual-cohort study quantifies the frequency of perineural invasion, invasion below the dermis, lack of cellular differentiation, and large tumor diameter in squamous cell carcinomas in organ transplant recipients and the general population.
Key Points
Question
Are high-risk tumor features more common in squamous cell carcinomas (SCCs) in organ transplant recipients (OTRs) than in SCCs in the general population?
Findings
In this dual-cohort study comparing 741 SCCs from 191 OTRs with 2558 SCCs from 1507 population-based persons, with age and sex adjustment, perineural invasion and invasion to or beyond subcutaneous fat were more than twice as common in OTRs than in the general population. Poorly vs well-differentiated SCCs were 3-fold more common, and prevalence of SCCs greater than 20 mm vs 20 mm or smaller was moderately higher, with all differences significant.
Meaning
Squamous cell carcinomas in OTRs warrant intensive surveillance, preferably by multidisciplinary specialists.
Abstract
Importance
The extent to which major high-risk features of squamous cell carcinomas (SCCs) in organ transplant recipients (OTRs) differ from SCCs in the general population is not known.
Objective
To quantify the relative frequency of perineural invasion, invasion below the dermis, lack of cellular differentiation, and tumor diameter greater than 20 mm in SCCs in OTRs and the general population, by anatomic site.
Design, Setting, and Participants
This dual-cohort study in Queensland, Australia, included a cohort of OTRs at high risk of skin cancer ascertained from 2012 to 2015 (Skin Tumours in Allograft Recipients [STAR] study) and a population-based cohort ascertained from 2011 (QSkin Sun and Health Study). The STAR study comprised population-based lung transplant recipients and kidney and liver transplant recipients at high risk of skin cancer recruited from tertiary centers and diagnosed with histopathologically confirmed SCC from 2012 to 2015. The QSkin participants were recruited from Queensland’s general adult population, and primary SCCs diagnosed from 2012 to 2015 were ascertained through Medicare (national health insurance scheme) and linked with histopathology records. Data analysis was performed from July 2022 to April 2023.
Main Outcomes and Measures
Prevalence ratio (PR) of head/neck location, perineural invasion, tumor invasion to/beyond subcutaneous fat, poor cellular differentiation, and tumor diameter greater than 20 mm among SCCs in OTRs vs the general population.
Results
There were 741 SCCs excised from 191 OTRs (median [IQR] age, 62.7 [56.7-67.1] years; 149 [78.0%] male) and 2558 SCCs from 1507 persons in the general population (median [IQR] age, 63.7 [58.0-68.8] years; 955 [63.4%] male). The SCCs developed most frequently on the head/neck in OTRs (285, 38.6%), but on arms/hands in the general population (896, 35.2%) (P < .001). After adjusting for age and sex, perineural invasion was more than twice as common in OTRs as in population cases (PR, 2.37; 95% CI, 1.70-3.30), as was invasion to/beyond subcutaneous fat (PR, 2.37; 95% CI, 1.78-3.14). Poorly vs well-differentiated SCCs were more than 3-fold more common in OTRs (PR, 3.45; 95% CI, 2.53-4.71), and prevalence of tumors greater than 20 mm vs 20 mm or smaller was moderately higher in OTRs (PR, 1.52; 95% CI, 1.08-2.12).
Conclusions and Relevance
In this dual-cohort study, SCCs in OTRs had significantly worse prognostic features than SCCs in the general population, reinforcing the necessity of early diagnosis and definitive management of SCCs in OTRs.
Introduction
Squamous cell carcinomas (SCCs) of the skin develop up to 77 times more frequently in immunosuppressed organ transplant recipients (OTRs) than the general population.1,2 Because SCCs cause substantially more morbidity and death in the former, they are postulated to be innately more aggressive than in immunocompetent patients,2 but OTRs’ higher SCC mortality may simply reflect greater SCC tumor burdens per patient.3 The few studies that have compared histopathologic features of SCCs in transplant recipients and immunocompetent patients have variously reported differences in depth of invasion, perineural or lymphatic invasion, and degree of differentiation4,5,6 consistent with more aggressive SCCs in OTRs. However, no studies have directly compared all 5 core clinicopathologic indicators of poor SCC outcomes, namely cephalic location, perineural invasion, invasion (to/beyond subcutaneous fat), poor differentiation, and tumor size greater than 20 mm7 in each group. We therefore aimed to quantify the relative prevalence of these standard prognostic features in SCCs excised from OTRs vs the general population in Queensland, Australia, while accounting for association of SCC features in patients with multiple tumors.
Methods
Information about SCCs treated by wide local excision was collected during 2 prospective studies: Skin Tumours in Allograft Recipients (STAR) study, and QSkin Sun and Health Study, with institutional ethics committees’ approval and all participants’ written informed consent. STAR participants recruited from tertiary centers comprised a population-based sample of consecutive lung transplant recipients,1 while kidney and liver transplant recipients were selected for their high risk of skin cancer based on previous disease, or age 40 years or older or receiving a transplant 10 years or more prior (thereby capturing almost all SCC cases).8 All those diagnosed from 2012 to 2015 with 1 or more histopathologically confirmed SCCs were included. The QSkin participants were recruited in 2011 from Queensland’s adult population via the state electoral roll, and primary skin cancer treatments ascertained through Medicare (Australia’s universal health insurance agency) and linked with histopathology records.9 All primary SCCs diagnosed from 2012 to 2015 were included.
Details of anatomic site, perineural invasion, level of tissue invasion, degree of cellular differentiation, and tumor diameter were extracted from histopathology reports. Frequency distributions of SCC features in OTRs and the general population were compared using χ2 tests. We estimated the prevalence ratios (PRs) for SCC features in OTRs vs the general population using a log binomial regression model with binary outcome, adjusting for sex and baseline age. Generalized estimating equations were used to adjust for intraindividual correlations among multiple SCCs in individuals. All statistical analyses were performed using SAS, version 9.4 (SAS Institute). This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Results
We compared 741 SCCs excised from 191 OTRs with 2558 SCCs from 1507 non-OTRs in the general population. Median (IQR) ages at diagnosis of first SCC during follow-up were 62.7 (56.7-67.1) years (OTRs) and 63.7 (58.0-68.8) years (population cases); male patients comprised 78.0% (149 of 191) and 63.4% (955 of 1507), respectively (Table 1). In OTRs, SCCs developed most frequently on the head/neck (285, 38.6%), but on arms/hands in the general population (896, 35.2%) (P < .001) (Table 2). Among head/neck SCCs, 16.8% were on the scalp in OTRs compared with 8.6% in the general population. Perineural invasion was more common in OTRs than population cases (3.0% vs 0.9%), as was SCC invasion to/beyond subcutaneous fat (11.1% vs 3.1%) (both P < .001). The SCCs in OTRs were poorly differentiated more often than in the general population (10.3% vs 2.3%) (P < .001) and more likely to be greater than 20 mm in diameter (4.8% vs 2.6%) (P = .002) (Table 2). In both populations, poor prognostic features of perineural invasion, invasion to/beyond subcutaneous fat, and lack of differentiation occurred more often in SCCs on the head/neck than other sites, especially tumor invasion, while the reverse was true for tumor diameter greater than 20 mm (eTable 1 in Supplement 1). Within OTRs (with median [range] of 2 [1-39] SCCs per patient), poorer prognostic features were overrepresented in the lung transplant cohort (eTable 2 in Supplement 1), while duration of immunosuppression was not associated.
Table 1. Demographic Characteristics of Patients With SCCs in Organ Transplant Recipients and the General Population.
| Characteristic | No. (%) | χ2 P value | |
|---|---|---|---|
| Organ transplant recipients (n = 191) | General population (n = 1507) | ||
| Age at diagnosis,a mean (SD); median (IQR), y | 61.4 (9.1); 62.7 (56.7-67.1) | 62.4 (6.8); 63.7 (58.0-68.8) | .16b |
| Age at diagnosis,a y | |||
| <50 | 16 (8.4) | 88 (5.8) | .39 |
| 50-59 | 50 (26.2) | 409 (27.1) | |
| 60-69 | 100 (52.4) | 846 (56.1) | |
| ≥70 | 25 (13.1) | 164 (10.9) | |
| Sex | |||
| Female | 42 (22.0) | 552 (36.6) | <.001 |
| Male | 149 (78.0) | 955 (63.4) | |
| Prior skin cancerc | |||
| No history of skin cancer | 14 (7.9) | 94 (6.3) | .40 |
| At least 1 treatment of skin cancer | 163 (92.1) | 1405 (93.7) | |
| Skin checks by physicianc | |||
| None or once a year | 56 (31.6) | 457 (30.8) | <.001 |
| 2-5 Times in the past 3 y | 27 (15.3) | 635 (42.7) | |
| >5 Times in the past 3 y | 94 (53.1) | 394 (26.5) | |
| Transplant type | |||
| Kidney | 107 (56.0) | NA | NA |
| Lung | 30 (15.7) | NA | |
| Liver | 54 (28.3) | NA | |
| Duration of immunosuppressionc | |||
| <5 y | 56 (29.3) | NA | NA |
| 5-10 y | 57 (29.8) | NA | |
| >10 y | 78 (40.8) | NA | |
Abbreviations: NA, not applicable; SCC, squamous cell carcinoma.
First SCC after study entry.
T test comparing mean age between 2 groups.
At study entry; totals may not sum to overall total due to missing values.
Table 2. Tumor Characteristics of SCCs in Organ Transplant Recipients and the General Population.
| Characteristic | No. (%) | P value | |
|---|---|---|---|
| Organ transplant recipients (n = 741) | General population (n = 2558) | ||
| Body site | |||
| Head/neck | 285 (38.6) | 812 (31.9) | <.001 |
| Scalp | 48 (16.8) | 70 (8.6) | |
| Ear | 35 (12.3) | 102 (12.6) | |
| Face | 148 (51.9) | 512 (63.1) | |
| Lip | 19 (6.7) | 33 (4.1) | |
| Neck | 35 (12.3) | 95 (11.7) | |
| Trunk | 78 (10.6) | 237 (9.3) | |
| Arms/hands | 240 (32.5) | 896 (35.2) | |
| Legs/feet | 135 (18.3) | 602 (23.6) | |
| Missing | 3 | 11 | NA |
| Perineural invasion | |||
| No | 542 (97.0) | 1797 (99.1) | <.001 |
| Yes | 17 (3.0) | 16 (0.9) | |
| Missing | 182 | 745 | NA |
| Tumor invasion | |||
| Dermis | 499 (88.9) | 2255 (96.9) | <.001 |
| To/beyond subcutaneous fat | 62 (11.1) | 73 (3.1) | |
| Missing | 180 | 230 | NA |
| Tumor grade | |||
| Poorly differentiated | 73 (10.3) | 56 (2.3) | <.001 |
| Moderately differentiated | 343 (48.5) | 772 (31.6) | |
| Well differentiated | 291 (41.2) | 1617 (66.1) | |
| Missing | 34 | 113 | NA |
| Tumor size | |||
| ≤20 mm | 654 (95.2) | 2404 (97.5) | .002 |
| >20 mm | 33 (4.8) | 63 (2.6) | |
| Missing | 54 | 91 | NA |
Abbreviations: NA, not applicable; SCC, squamous cell carcinoma.
After adjusting for age and sex, perineural invasion by head/neck SCCs was more than twice as common in OTRs than the population (PR, 2.37; 95% CI, 1.70-3.30) and more than 4 times as common in SCCs on the arms/hands in OTRs (PR, 4.51; 95% CI, 1.15-17.66), but with no difference between groups for perineural invasion among trunk/lower limb SCCs (Table 3). Invasion to/beyond subcutaneous fat was more than twice as common for SCCs in OTRs overall (PR, 2.37; 95% CI, 1.78-3.14) and for SCCs on head/neck and trunk/lower limbs, but not arms/hands. Very large differences were seen in prevalence of poorly vs well-differentiated SCCs in OTRs compared with the general population overall (PR, 3.45; 95% CI, 2.53-4.71), and for SCCs on arms/hands especially (PR, 4.75; 95% CI, 3.20-7.05). Prevalence of tumors measuring greater than 20 mm vs 20 mm or smaller was moderately higher in OTRs, both overall (PR, 1.52; 95% CI, 1.08-2.12) and on specific sites (Table 3).
Table 3. Age-Adjusted and Sex-Adjusted Prevalence Ratios for Comparison of SCC Tumors Overall and by Site in Organ Transplant Recipients and the General Population.
| Prevalence ratio (95% CI) | ||||
|---|---|---|---|---|
| All | Head/neck | Arms/hands | Trunk/lower limb | |
| Body site | ||||
| Head/neck | 1.15 (0.86-1.56) | NA | NA | NA |
| Arms/hands | 1.05 (0.82-1.35) | NA | NA | NA |
| Trunk/lower limb | 1 [Reference] | NA | NA | NA |
| Perineural invasion | ||||
| Yes | 2.37 (1.70-3.30) | 2.55 (1.87-3.48) | 4.51 (1.15-17.66) | 0.80 (0.14-4.58) |
| No | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Level of tumor invasion | ||||
| Dermis | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| To/beyond subcutaneous fat | 2.37 (1.78-3.14) | 2.44 (1.83-3.24) | 1.53 (0.67-3.49) | 2.70 (1.37-5.34) |
| Tumor grade | ||||
| Well differentiated | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Moderately differentiated | 1.89 (1.51-2.36) | 1.68 (1.23-2.31) | 2.30 (1.76-2.99) | 1.83 (1.29-2.62) |
| Poorly differentiated | 3.45 (2.53-4.71) | 3.14 (2.17-4.53) | 4.75 (3.20-7.05) | 3.09 (1.77-5.37) |
| Tumor size | ||||
| ≤20 mm | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| >20 mm | 1.52 (1.08-2.12) | 1.62 (1.17-2.24) | 1.24 (0.70-2.19) | 2.07 (1.34-3.21) |
Abbreviations: NA, not applicable; SCC, squamous cell carcinoma.
Discussion
We have shown that, in general, the cardinal (rather than emerging10) histopathologic features associated with poor prognosis are more prevalent in SCCs in OTRs than in the general population. Largest differences between SCCs in OTRs and the general population were seen for SCCs on arms/hands, where perineural invasion and poor differentiation of tumor tissue were more than 4 times more frequent in OTRs, although there was no difference in level of SCC invasion or frequency of SCCs greater than 20 mm on this site. Features of head/neck SCCs were uniformly more aggressive in OTRs, with perineural invasion, invasion to/beyond subcutaneous fat, poor differentiation, and tumor diameter greater than 20 mm all more than twice as frequent as among SCCs arising in the general population. While SCCs on the less common sites of trunk/lower limbs showed no difference in frequency of perineural invasion between OTRs and the general population, other histologic indicators of poor prognosis occurred significantly more often in trunk/limb SCCs in OTRs.
These findings are considered generalizable, confirming that OTRs’ poorer SCC outcomes are associated with not only their sheer numbers of SCC tumors,3 but also with a strong shift toward more invasive, less differentiated, and larger SCC tumors, in agreement with previous findings.4,5,6
This shift is likely associated with decreased immunosurveillance resulting from immunosuppressive therapy (since carcinogenesis decelerates with therapy cessation11) interacting with effects of high UV radiation exposure.1 Our findings extend knowledge of site-specific differences in prognostic features of SCCs, with head/neck SCCs in OTRs showing higher prevalence of all major indicators of poor prognosis than in the general population. They reinforce the necessity of early diagnosis of SCCs in OTRs through frequent surveillance, including of the less frequently affected trunk/lower limbs, since we found that trunk/lower limb SCCs are more likely to be poorly differentiated, invading below the dermis when diagnosed (and likely more fast-growing) compared with SCCs in the general population.
Limitations
The main study limitation was the lack of central review of SCCs to ensure standard assessment of histopathologic features, including caliber of nerves with perineural invasion and cell differentiation; such a review would not have been feasible logistically. Its strengths were the comprehensive comparison of major prognostic features of SCCs that were ascertained prospectively in cohort studies of OTRs and population-based cases, and the accounting for multiple SCCs per case during analysis.
Conclusions
This dual-cohort study shows that SCCs in OTRs possess significantly worse high-risk histopathologic features than SCCs in the general population. This finding, along with the greater multiplicity of SCCs in OTRs,3 warrants intensive surveillance of affected OTRs optimally with multidisciplinary specialist care, at the same time recognizing that all SCCs with high-risk features warrant intensive therapy and follow-up.
eTable 1. Demographic and tumor characteristics for SCC tumors in organ transplant recipients and general population cohorts stratified by site
eTable 2. Characteristics of SCC tumors by transplant type
Data Sharing Statement
References
- 1.Way M, Marquart L, Chambers DC, et al. Skin cancer multiplicity in lung transplant recipients: a prospective population-based study. Br J Dermatol. 2020;183(3):503-508. doi: 10.1111/bjd.18812 [DOI] [PubMed] [Google Scholar]
- 2.Euvrard S, Kanitakis J, Claudy A. Skin cancers after organ transplantation. N Engl J Med. 2003;348(17):1681-1691. doi: 10.1056/NEJMra022137 [DOI] [PubMed] [Google Scholar]
- 3.Gonzalez JL, Reddy ND, Cunningham K, Silverman R, Madan E, Nguyen BM. Multiple cutaneous squamous cell carcinoma in immunosuppressed vs immunocompetent patients. JAMA Dermatol. 2019;155(5):625-627. doi: 10.1001/jamadermatol.2018.5595 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Smith KJ, Hamza S, Skelton H. Histologic features in primary cutaneous squamous cell carcinomas in immunocompromised patients focusing on organ transplant patients. Dermatol Surg. 2004;30(4, pt 2):634-641. doi: 10.1111/j.1524-4725.2004.30149.x [DOI] [PubMed] [Google Scholar]
- 5.Harwood CA, Proby CM, McGregor JM, Sheaff MT, Leigh IM, Cerio R. Clinicopathologic features of skin cancer in organ transplant recipients: a retrospective case-control series. J Am Acad Dermatol. 2006;54(2):290-300. doi: 10.1016/j.jaad.2005.10.049 [DOI] [PubMed] [Google Scholar]
- 6.Lott DG, Manz R, Koch C, Lorenz RR. Aggressive behavior of nonmelanotic skin cancers in solid organ transplant recipients. Transplantation. 2010;90(6):683-687. doi: 10.1097/TP.0b013e3181ec7228 [DOI] [PubMed] [Google Scholar]
- 7.Farberg AS, Fitzgerald AL, Ibrahim SF, et al. Current methods and caveats to risk factor assessment in cutaneous squamous cell carcinoma (cSCC): a narrative review. Dermatol Ther (Heidelb). 2022;12(2):267-284. doi: 10.1007/s13555-021-00673-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Iannacone MR, Sinnya S, Pandeya N, et al. ; STAR Study . Prevalence of skin cancer and related skin tumors in high-risk kidney and liver transplant recipients in Queensland, Australia. J Invest Dermatol. 2016;136(7):1382-1386. doi: 10.1016/j.jid.2016.02.804 [DOI] [PubMed] [Google Scholar]
- 9.Olsen CM, Green AC, Neale RE, et al. ; QSkin Study . Cohort profile: the QSkin Sun and Health Study. Int J Epidemiol. 2012;41(4):929-929i. doi: 10.1093/ije/dys107 [DOI] [PubMed] [Google Scholar]
- 10.Zakhem GA, Pulavarty AN, Carucci J, Stevenson ML. Association of patient risk factors, tumor characteristics, and treatment modality with poor outcomes in primary cutaneous squamous cell carcinoma: a systematic review and meta-analysis. JAMA Dermatol. 2023;159(2):160-171. doi: 10.1001/jamadermatol.2022.5508 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Otley CC, Maragh SL. Reduction of immunosuppression for transplant-associated skin cancer: rationale and evidence of efficacy. Dermatol Surg. 2005;31(2):163-168. doi: 10.1097/00042728-200502000-00008 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Demographic and tumor characteristics for SCC tumors in organ transplant recipients and general population cohorts stratified by site
eTable 2. Characteristics of SCC tumors by transplant type
Data Sharing Statement