To the Editor: Cutaneous squamous cell carcinoma (CSCC) is the second most frequent cancer in humans, and it can be both locally invasive and metastatic to distant sites. Despite distant metastases (DM) being rare in CSCC, some patients do develop DM. Immunosuppression (IS) is a risk factor for CSCC. Several clinical practice guidelines consider IS to be a high-risk feature for recurrence in CSCC, but it is currently excluded from most popular staging systems. Some recent studies, including a meta-analysis,1 have explored the prognostic impact of IS in CSCC.1,2 However, there is still certain lack of knowledge on IS impact on DM as a separate outcome instead of pooling it with nodal metastases given the very small number of studies with small series in which DM was evaluated.1
We evaluated a cohort of patients diagnosed with CSCC between 2010 and 2019 at the University Hospital of Salamanca, Spain. CSCCs were selected when 1 or more of the following risk factors of poor prognosis were present: tumor diameter≥2 cm, thickness>6 mm, perineural invasion, poor differentiation, lymphovascular invasion (LVI), invasion beyond the subcutaneous fat, desmoplasia, tumor budding, and IS. When more than 1 CSCC was diagnosed in the same patient, that supposing the greatest risk was considered. A cohort of 781 high-risk CSCCs was retrieved (218 in immunosuppressed). In this cohort, there were 100 local recurrences, 97 nodal metastases, 30 DM, and 64 patients died from CSCC.
We first compared CSCCs between immunocompetent and immunosuppressed patients (Table I). Subsequently, we explored the cumulative incidence function using the Gray test, considering death from other causes as a competing risk. Multivariate models, using Fine–Gray proportional hazard regression, were derived to assess the independence of risk factors in CSCC outcomes. Specifically, we considered BWH T-stage, LVI (relevant in CSCC prognosis3 despite not considered for staging), gender,1,4 and IS. We found that LVI and the T-stage were consistently associated with all poor outcome events. IS was associated with distant metastases in the univariate (HR = 2.662 [95% CI = 1.303-5.439], P = .0072) and multivariate (HR = 3.393 [95% CI = 1.646-6.995], P = .00093) analyses. Gender was not associated with prognosis in this study (Table II).
Table I.
Characteristics of cutaneous squamous cell carcinomas developed in immunocompetent (group 1) and immunocompromised (group 2) patients and the cause of immunosuppression in these cases
| Group 1 Inmunocompetent (N = 563) |
Group 2 Immunosuppressed (N = 218) |
P value | |
|---|---|---|---|
| Patient history | |||
| Age, Me (IQR) | 86 (10) | 82 (12) | .01 |
| Sex | |||
| Male | 280 (49.7 %) | 142 (65.1 %) | .01 |
| Female | 283 (50.3 %) | 76 (34.9 %) | |
| Tumor traits | |||
| Tumor size > 4 cm | |||
| > 2 cm | 348 (61.9 %) | 145 (66.5 %) | N.S. |
| [2-4) cm | 175 (31.1 %) | 56 (25.7 %) | |
| ≥ 4 cm | 39 (6.9 %) | 17 (7.8 %) | |
| Tumor thickness | |||
| Up to 6 mm | 311 (55.3 %) | 126 (57.8 %) | N.S. |
| ≥ 6 mm | 251 (44.7 %) | 92 (42.2 %) | |
| Tumor invasion | |||
| Beyond fat | 140 (24.9 %) | 60 (27.5 %) | N.S. |
| Muscle | 101 (18 %) | 36 (16.5 %) | |
| Bone | 7 (3.4 %) | 6 (9.7 %) | |
| Tumor location | |||
| H&N High risk | 185 (33 %) | 76 (34.9 %) | .008 |
| H&N Low risk | 150 (26.7 %) | 80 (36.7 %) | |
| Trunk and Limbs | 113 (20.2 %) | 29 (13.3 %) | |
| Hand & Feet | 109 (19.5 %) | 30 (13.8 %) | |
| Grade of differentiation | |||
| Good | 97 (17.3 %) | 44 (20.3 %) | N.S. |
| Moderate | 343 (61 %) | 122 (56.2 %) | |
| Poor | 122 (21.7 %) | 51 (23.5 %) | |
| Lymphovascular invasion | |||
| Yes | 18 (3.2 %) | 4 (1.8 %) | N.S. |
| No | 545 (96.8 %) | 214 (98.2 %) | |
| Perineural invasion | |||
| No | 419 (74.6 %) | 172 (78.9 %) | N.S. |
| Nerve ≤ 0.1 mm | 69 (12.3 %) | 25 (11.5 %) | |
| Nerve > 0.1 mm | 74 (13.2 %) | 21 (9.6 %) | |
| Other cause | 342 (88.1 %) | 129 (87.8 %) | |
| Follow-up months, Me (IQR) | 40 (58) | 31.5 (40) | .005 |
| Staging | |||
| 8th-AJCC | |||
| T1-T2 | 256 (45.6 %) | 101 (46.3 %) | N.S. |
| T3-T4 | 306 (54.4 %) | 117 (53.7 %) | |
| BWH | |||
| T1-T2a | 408 (72.6 %) | 164 (75.2 %) | N.S. |
| T2b-T3 | 154 (27.4 %) | 54 (24.8 %) |
| Immunosuppression type | n (%) |
|---|---|
| Chronic lymphocytic leukemia | 30 (13.7) |
| Other hematological malignancies | 51 (23.4) |
| Solid organ transplantation | 14 (6.4) |
| Chronic immunosuppressive treatment | 5 (2.3) |
| Poorly controlled diabetes | 32 (14.7) |
| Chronic renal failure | 22 (10.1) |
| Solid tumors (excluding nonmelanoma skin cancer) | 49 (22.5) |
| Inflammatory disease | 18 (8.2) |
| HIV | 2 (0.9) |
Proportions were compared using the chi-squared test (using IBM SPSS Statistics v26). Bold values of P < .05 were considered statistically significant.
AJCC, The American Joint Committee on Cancer; BWH, Brigham and Women's Hospital; H&N, head and neck; HIV, human immunodeficiency virus; IQR, interquartile range; Me, median; NS, not significant.
Table II.
Univariate and multivariate Fine-Gray regression analyses to assess the degree of association of immunosuppression, along with other tumor variables, to various adverse outcomes of cutaneous squamous cell carcinoma
| Outcome | Variable | Univariate model |
Multivariate model |
||||
|---|---|---|---|---|---|---|---|
| HR | IC (95 %) | P | HR | IC (95 %) | P | ||
| Local recurrence | Gender | 1.256 | 0.8443-1.869 | .26 | 1.155 | 0.7734-1.726 | .48 |
| Immunosuppression | 1.233 | 0.812-1.871 | .33 | 1.239 | 0.8144-1.885 | .32 | |
| Lymphovascular invasion | 2.419 | 1.196-4.892 | .014 | 1.998 | 0.9187-4.346 | .081 | |
| BWH (T2b/T3 vs T1/T2a) | 3.018 | 2.043-4.458 | <.0001 | 2.87 | 1.94-4.244 | <.0001 | |
| Nodal metastases | Gender | 0.9204 | 0.6195-1.367 | .68 | 0.9022 | 0.6004-1.356 | .62 |
| Immunosuppression | 0.9465 | 0.6043-1.482 | .81 | 1.08 | 0.6832-1.707 | .74 | |
| Lymphovascular invasion | 5.307 | 2.822-9.978 | <.0001 | 4.833 | 2.926-7.982 | <.0001 | |
| BWH (T2b/T3 vs T1/T2a) | 4.776 | 3.185-7.163 | <.0001 | 4.282 | 2.859-6.415 | <.0001 | |
| Distant metastases | Gender | 0.7545 | 0.3696-1.54 | .44 | 0.6738 | 0.3082-1.473 | .32 |
| Immunosuppression | 2.662 | 1.303-5.439 | .0072 | 3.393 | 1.646-6.995 | .00093 | |
| Lymphovascular invasion | 6.003 | 2.669-13.5 | <.0001 | 5.808 | 2.349-14.36 | .00014 | |
| BWH (T2b/T3 vs T1/T2a) | 4.229 | 2.04-8.767 | .00011 | 3.756 | 1.792-7.874 | .00046 | |
| Disease-specific death | Gender | 1.34 | 0.7866-2.281 | .28 | 0.6758 | 0.3938-1.16 | .16 |
| Immunosuppression | 1.003 | 0.5545-1.816 | .99 | 1.129 | 0.6181-2.063 | .69 | |
| Lymphovascular invasion | 5.381 | 2.679-10.81 | <.0001 | 3.596 | 1.799-7.189 | .00029 | |
| BWH (T2b/T3 vs T1/T2a) | 7.205 | 4.03-12.88 | <.0001 | 6.692 | 3.76-11.91 | <.0001 | |
We decided to use BWH T-stage (which encompasses several high-risk features including perineural invasion ≥0.1 mm, poor degree of differentiation, invasion beyond the subcutaneous fat and tumor size ≥2 cm), lymphovascular invasion (known to be relevant in Cutaneous squamous cell carcinomaprognosis despite not being considered in staging systems), male sex (another feature recently associated with poor prognosis in cohort and population-based studies) and immunosuppression. All these tests were carried out using R (version 4.3.0).
Statistically significant P values (<.05) are in bold.|
BWH, Brigham and Women's Hospital.
IS is unanimously considered a high-risk feature in CSCC, but it needs a more homogeneous definition given that different levels and types of IS might have different effects on prognosis.1 While recent papers failed to identify differences in prognosis between immunosuppressed and immunocompetent patients when adjusted by T-stage,1 all metastases were pooled together rather than splitting them by type. A recent meta-analysis demonstrated IS to be a risk factor for all poor outcome events in CSCC except for DM, but only 2 studies had explored the impact of IS in DM development.2 Few studies have focused on DM in CSCC.5 Here, we found that IS is a patient feature independently associated with a higher risk of DM, despite there was not an association when all metastases were pooled nor with nodal metastases separately. Further studies will help define IS accurately and identify its genuine prognostic impact in CSCC and if DM development exhibits specific features compared with other outcome events.
Conflicts of interest
None disclosed.
Footnotes
Funding sources: Javier Cañueto is partially supported by the (Gerencia Regional de Salud de Castilla y León, grants GRS2338/A/21 and GRS2549/A/22), by the Instituto de Salud Carlos III (ISCIII), (PI18/00587 and PI21/01207), co-financed by European Union, and by the “Programa de Intensificación of the ISCIII, grant number INT20/00074”. Sponsors and funders had no influence on the design and conduct of the study, the collection, management, analysis, and interpretation of the data, the preparation, review, and approval of the manuscript, or the decision to submit the manuscript for publication.
Patient consent: Not applicable.
IRB approval status: Not applicable.
References
- 1.O'Connor D.M., Murad F., Danesh M.J., et al. Immune status does not independently influence cutaneous squamous cell carcinoma metastasis and death when stratified by tumor stage: a dual-center retrospective cohort analysis of primary N0 disease. J Am Acad Dermatol. 2022;87:1295–1302. doi: 10.1016/j.jaad.2022.08.050. [DOI] [PubMed] [Google Scholar]
- 2.Zakhem G.A., Pulavarty A.N., Carucci J., Stevenson M.L. 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;159:160–171. doi: 10.1001/jamadermatol.2022.5508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kus K.J.B., Murad F., Smile T.D., et al. Higher metastasis and death rates in cutaneous squamous cell carcinomas with lymphovascular invasion. J Am Acad Dermatol. 2022;86:766–773. doi: 10.1016/j.jaad.2021.11.002. [DOI] [PubMed] [Google Scholar]
- 4.Tokez S., Venables Z.C., Hollestein L.M., et al. Risk factors for metastatic cutaneous squamous cell carcinoma: refinement and replication based on 2 nationwide nested case-control studies. J Am Acad Dermatol. 2022;87:64–71. doi: 10.1016/j.jaad.2022.02.056. [DOI] [PubMed] [Google Scholar]
- 5.Stefanovic N., Fitzmaurice C.J., Ormond P., Irvine A.D., Barry R.B. Risk factors for distant metastasis in cutaneous squamous cell carcinoma. Br J Dermatol. 2022;187:435–436. doi: 10.1111/bjd.21250. [DOI] [PMC free article] [PubMed] [Google Scholar]