Mohs micrographic surgery for keratinocyte carcinomas: clinicopathological predictors of the number of stages

Abstract

Background

The number of Mohs stages needed to remove a keratinocyte carcinoma affects resource use, expenses, and repair complexity. This study aimed to identify clinicopathological predictors associated with increased or decreased stages and areas for further research.

Methods

A retrospective review was conducted from a single private practice with two Mohs surgeons of 2788 consecutive Mohs cases between January 2017 and December 2021, analyzing the average number of stages taken versus national norms (P = 0.21) and subgroups using unpaired t tests (*<0.05).

Results

Several tumor features were significantly associated with fewer stages: squamous cell carcinomas, Mohs appropriate use criteria score of 7 or 8, preoperative size <0.25 cm², tumors on the lips and extremities (including hands/fingers), and smoking. Clinicopathological features significantly associated with more stages included Mohs appropriate use criteria score of 9, recurrent skin cancers, basal cell carcinomas, tumor size of 2.25–3.99 cm², cancers on ears, solid organ transplant patients, treatment delays >180 days, and patients ≥90 years old.

Conclusions

Significant predictors exist for both increased and decreased numbers of Mohs micrographic surgery stages required to eradicate a tumor, which may help Mohs surgeons facilitate, plan, and allocate resources more effectively. Areas for further research in Mohs micrographic surgery are identified.

Keratinocyte carcinomas (KCs) include basal cell carcinomas (BCCs) and squamous cell carcinomas (SCCs).¹ Multiple treatment options exist for these tumors, but Mohs micrographic surgery (MMS) is considered the “gold standard” to remove difficult-to-treat KCs.^2–5 MMS is performed in a stepwise method typically referred to as stages. Predicting the number of stages needed for MMS removal of KCs can be helpful in preoperative planning, including estimating the time and resources needed and the type of repair that will be necessary, and can give patients more accurate expectations for their surgery.^6–9 The purpose of this review was to identify clinicopathological predictors associated with an increased or decreased number of stages required to completely eradicate a KC utilizing MMS as well as identify areas where further research is needed.

Methods

Data on tumor features and patient characteristics were searched and retrieved from a proprietary electronic health record system as well as a Mohs surgical note generating database from January 1, 2017, through December 31, 2021, for tumors treated with MMS. A total of 11,264 consecutive KCs were diagnosed on 5878 patients, with MMS performed on 2788 (24.8%) tumors during this timeframe. Every tumor treated with MMS was included in the review. Each tumor had a preoperative biopsy, either in house or via outside referral. MMS was performed in the typical fashion by two private-practice board-certified dermatologists (dermatology and micrographic dermatological surgery) with curettage¹⁰ or scalpel debulking, as indicated, before the first stage, and taking 1 to 3 mm margins for each stage based on the surgeon’s judgment.^4,5,11,12 BCCs were stratified into nodular, micronodular, infiltrative, sclerosing, superficial, other, or unspecified, as well as mixed aggressive and mixed nonaggressive. SCCs were stratified into moderately differentiated, well-differentiated including keratoacanthoma type, poorly differentiated, acantholytic, superficial, basosquamous, in situ, and other or unspecified.

Unpaired t tests were performed via Excel comparing the average number of stages for all nonmelanoma skin cancer Mohs cases in our data set (X = 1.77, N = 2788, SD = 0.95) versus subgroups. Our data set was statistically insignificant when compared to national norms (P = 0.21) previously established by Krishnan et al regarding average number of stages taken from 17,311 cases by 1845 Mohs surgeons between 2012 and 2014 (X = 1.74, SD = 0.44).^13,14

Results

For the 2788 MMS cases, the average number of stages required for tumor removal was 1.77. All subgroups were compared to the average rather than to one another to establish significance, as shown in Table 1. Overall, BCC required significantly more stages than SCC (1.84 vs. 1.57). A Mohs appropriate use criteria (AUC) score of 9 was significantly associated with more stages required for tumor removal at 1.92 stages (P < 0.001) compared to the average. An AUC score of 7 or 8 was associated with fewer stages at 1.51 and 1.68, respectively (P < 0.001, P = 0.004) compared to the average. Recurrent skin cancers required more stages than nonrecurrent skin cancers (1.98 vs. 1.75, P = 0.003).

Table 1.

Mohs stages based on clinical and demographic variables^*

Variable	N	%	Stages (n)	SD	P value
MMS KC overall
Total Mohs cases	2788		1.77	0.95
Basal cell carcinoma	2056	73.7	1.84	0.99	0.013*
Squamous cell carcinoma	732	26.3	1.57	0.79	<0.001*
Mohs appropriate use criteria
7, 8, or 9	2775	99.5	1.77	0.95
7	255	9.2	1.51	0.77	<0.001*
8	1295	46.7	1.68	0.90	0.004*
9	1225	44.1	1.92	1.00	<0.001*
Nature of KC
Recurrent	205	7.4	1.98	1.11	0.003*
Nonrecurrent	2583	92.6	1.75	0.93	0.436
Basal cell carcinoma
Nodular	899	43.7	1.82	1.01	0.177
Micronodular	160	7.8	1.84	0.90	0.364
Infiltrative	638	31.0	1.84	0.93	0.092
Sclerosing	57	2.8	1.91	1.01	0.271
Superficial	152	7.4	1.81	1.00	0.614
Other/unspecified	85	4.1	2.02	1.15	0.018*
Mixed	513	25.0	1.84	0.95
Aggressive	401	19.5	1.82	0.91	0.322
Nonaggressive	112	5.4	1.89	1.08	0.193
Squamous cell carcinoma
Moderate	331	45.2	1.63	0.82	0.010*
Well-differentiated/KA	198	27.0	1.56	0.64	0.002*
Poorly differentiated	6	0.8	1.67	0.75	0.797
Acantholytic	7	1.0	1.14	0.35	0.079
Superficial	113	15.4	1.44	0.85	<0.001*
Basosquamous	65	3.2	1.94	1.21	0.157
In situ	9	1.2	1.67	0.94	0.753
Other/unspecified	68	9.3	1.62	0.91	0.198
Preoperative sizes (cm2)
<0.25	243	8.7	1.61	0.86	0.011*
0.25–0.49	23.6	1.72	0.97	0.227
0.50–0.99	625	22.4	1.78	0.99	0.814
1.00–1.49	469	16.8	1.73	0.85	0.392
1.50–2.24	283	10.2	1.86	0.94	0.130
2.25–3.99	321	11.5	1.94	0.96	0.002*
4.00–8.99	146	5.2	1.81	0.94	0.620
≥9.00	44	1.6	1.84	1.11	0.630
Site-specific locations
Nose	827	29.7	1.84	1.02	0.067
Ears	510	18.3	1.91	0.94	0.002*
Eyelids	105	3.8	1.68	0.77	0.338
Lips	146	5.2	1.58	0.71	0.017*
Face, other	751	26.9	1.74	0.93	0.440
Scalp/neck	257	9.2	1.77	1.08	1.000
Upper extremities	32	1.1	1.34	0.47	0.011*
Hands/fingers	59	2.1	1.46	0.62	0.013*
Lower extremities/foot	87	3.1	1.45	0.64	0.002*
Trunk/chest	14	0.5	1.57	0.90	0.432
Delay from diagnosis to treatment (days)
0–29	2246	80.6	1.77	0.94	1.000
30–59	367	13.2	1.71	0.88	0.252
60–89	66	2.4	1.91	1.04	0.238
90–119	14	0.5	1.50	0.50	0.288
120–179	18	0.6	1.67	0.94	0.656
≥180	48	1.7	2.13	1.33	0.009*
Unknown	29	1.0	2.10	1.03	0.063
Age at diagnosis (years)
<20	1	0.0	3.00	0.00
20–39	18	0.6	1.94	1.13	0.449
40–59	281	10.1	1.73	1.04	0.505
60–6	513	18.4	1.72	0.89	0.269
70–79	972	34.9	1.76	0.93	0.776
80–89	813	29.2	1.79	0.94	0.597
≥90	190	6.8	1.93	1.01	0.025*
Transplant status
Solid organ transplant	49	1.8	2.06	1.30	0.036*
Nontransplant	2739	98.2	1.77	0.94	1.000
Insurance type
Medicare	1886	67.6	1.79	0.96	0.482
Medicaid	10	0.4	1.70	0.90	0.816
Commercial	672	24.1	1.71	0.93	0.140
Uninsured	220	7.9	1.79	0.89	0.763
Other patient characteristics
Male	1966	70.6	1.78	0.93	0.720
Female	822	29.4	1.74	0.99	0.431
Smoker	252	9.0	1.64	0.80	0.035*
Nonsmoker	2536	91.0	1.78	0.96	0.703
Urban ZIP code	2015	72.4	1.78	0.96	0.720
Rural ZIP code	768	27.6	1.75	0.91	0.602

*Total number of nonmelanoma skin cancers in the data set is 11,264 between January 1, 2017, and December 31, 2021. Significance (*) defined as <0.05.

KA indicates keratoacanthoma; KC: keratinocyte carcinoma.

Although not significant, for BCC, aggressive subtypes were associated with more stages, as compared to the average, with micronodular, infiltrating, and sclerosing requiring 1.84, 1.84, and 1.91 stages, respectively. Also, other/unspecified subtypes required significantly more stages at 2.00 (P = 0.02) compared to the average. The mixed types were divided into nonaggressive, which included both superficial and nodular subtypes, and aggressive, which included any other combination of subtypes. Although not significant, the mixed aggressive group required fewer stages than the mixed nonaggressive group at 1.82 and 1.89, respectively (P = 0.32, P = 0.19).

The subtype of SCC associated with the most stages compared to the average was basosquamous cell carcinoma at 1.94 stages, although this did not reach statistical significance (P = 0.16). The subtypes of SCC significantly associated with fewer stages, as compared to the average, included superficial, moderate, and well differentiated at 1.44, 1.63, and 1.56 stages, respectively (P < 0.001, P = 0.01, P = 0.002). The acantholytic subtype displayed the fewest number of stages at 1.14 stages (P = 0.08), as compared to the average, but did not reach statistical significance. A preoperative size of <0.25 cm² was significantly associated with a lower number of stages at 1.61 stages (P = 0.01) as compared to the average. Sizes of 2.25 to 3.99 cm² were significantly associated with an increased number of stages, at 1.94 (P = 0.002), as compared to the average.

The anatomical sites and their association with the number of stages is also noted in Table 1. The ears and the nose, as compared to the average, required an increased number of stages at 1.84 and 1.91, respectively, although only the ears reached statistical significance (P = 0.002). When compared to the average, the upper extremities, lower extremities, hands/fingers, and lips significantly required the least number of stages at 1.34, 1.45, 1.46, and 1.58 stages, respectively (P = 0.01, P = 0.002, P = 0.01, P = 0.02). The trunk was also associated with fewer stages at 1.57 as compared to the average.

Delays in treatment are shown in Table 1. Compared to the average number of stages required, an increased number of stages was seen with a delay of 60 to 89 days and >180 days at 1.91 and 2.13, respectively, with the latter reaching statistical significance (P = 0.009). Only one patient <20 years old was treated and required three stages. Otherwise, patients >90 years old had a significantly increased number of stages required at 1.91 (P = 0.03) as compared to the other age ranges.

As shown in Table 1, solid organ transplant patients required significantly more stages than nontransplant patients at 2.06 and 1.77, respectively (P = 0.04). No statistically significant differences were seen in the number of stages required for the insured versus the uninsured population. Finally, Table 1 shows no difference in the number of stages required between men and women or between urban and rural patients. However, positive smoking status was associated with a decrease in the number of stages, at 1.64 (P = 0.04).

Discussion

Predictors of the number of stages for MMS to completely eradicate a tumor have been described.^{6–10,15–22} In our data set, BCC represented 73.7% of KC tumors and it took more stages to remove a BCC at 1.84 versus an SCC at 1.57 stages, a difference also shown by Diel et al.²³ Compared to the BCC mean number of stages, the aggressive subtype of BCC sclerosing (morpheaform) required an increased number of stages for complete removal, as expected.^6,23,24 Traditionally, micronodular and infiltrative BCCs have been considered more aggressive^9,17; however, in our data set, the number of stages required for these BCCs was the same as the mean for all BCCs. The traditionally considered less aggressive BCCs of nodular and superficial⁷ required just under the mean number of stages for all BCCs. Of interest, 90 (4.4%) diagnosed BCCs had no subtype on the original pathology report, or one of the less common subtypes, and were not further categorized. The mean was 2.0 stages, indicating these tumors were likely one of the more aggressive subtypes and/or on patients with more concerning clinical characteristics. Although one would expect the mixed aggressive subtypes to take more stages, the mixed subsets required about the same number of stages as the mean of all BCCs.

SCCs accounted for 26.3% of the tumors in our data set. The number of stages required for moderately and well-differentiated SCCs was similar to the mean for all SCCs. Although not statistically significant, SCC in situ^25–27 and poorly differentiated SCC required slightly more stages than the mean for all SCCs. Superficial SCC required the least number of stages, except for acantholytic SCC. Acantholytic SCC has been previously considered an aggressive SCC, but this has been refuted.^28,29 Our data, although not statistically significant, showed it required fewer stages to remove this subtype of SCC compared to other subtypes of SCCs. The subtype basosquamous cell carcinoma was included under SCC because the biological behavior can be similar to SCC. Basosquamous cell carcinoma accounted for 3.2% of SCCs and required more than the average number of stages for SCC removal; however, this did not reach statistical significance. Finally, the number of stages for other subtypes of SCC and unspecified subtypes of SCC was similar to the mean for all SCCs.

The development of the Mohs AUC and implementation into clinical practice is to help define tumors that will benefit from MMS.^30–32 This score considers many clinicopathological factors, including tumor size and location, aggressiveness, and immunosuppression in patients, among other criteria.^6,30,31,33 A score of 7, 8, or 9 indicates appropriateness for MMS, and all were significant in our data set, with a score of 9 taking the most number of stages for complete tumor removal. Recurrent tumors required more stages for removal. This is likely because they often have more aggressive features than primary tumors.^23,24,34 In general, the larger preoperative lesion size of the tumor, the more stages required^{19–22,35,36} for eradication; however, the relationship was not completely linear.²³ Larger tumors tend to have more concerning histopathologic features.^19–22 Location of a KC on certain body parts demonstrated a relationship to the number of stages required, as has been previously reported.^{4,5,7,37–41} The most common location for KC that received MMS was the nose, at 29.7% of tumors; however, the ears, which accounted for 18.3% of tumors, required the greatest number of stages. Mulvany et al showed that BCCs of the ear were often more aggressive than on other head and neck locations.³⁹ Delay in treatment after diagnosis has been associated with larger defects but not necessarily with more difficult repairs.^21,35,42 In our data set, the mean number of days from diagnosis until treatment was 28. Delays >180 days required more stages for tumor removal, but there was not a linear relationship.

Patient age had a relationship with the number of stages required for tumor removal, with younger as well as older patients taking more stages.^3,30,43 Older patients typically have more concerning clinicopathological features, and a priority of making smaller scars in younger patients may have contributed to increased stages in these age groups. Likely because of immunosuppression, solid organ transplant patients⁴⁴ required significantly more stages to completely eradicate tumor compared to nontransplant patients. It is noted that the calcineurin inhibitors, cyclosporine A and tacrolimus, greatly increase the risk of KC, particularly SCC.^45,46 Tacrolimus, because of its side effect profile, has recently been utilized more than cyclosporine A for immunosuppression in transplant patients.^37,41,45,46 However, the mTOR inhibitors, sirolimus and everolimus, which are also effective immunosuppressants, may actually be protective against KC and other cancers.^30,45–47 Insurance status seemed to have very little influence on the number of stages required for tumor removal.^48,49 The number of stages required for tumor removal was insignificant between men and women. Only 9.0% of our population indicated they smoked; however, smokers required fewer stages than nonsmokers.⁵⁰ SCC has been noted to be more common in smokers, especially of the lips and oral cavity.⁵¹ Whether the patient, at the time of surgery, lived in urban versus rural zip codes (adjusted proportionally for our database) had no significant influence on the number of stages required for tumor removal.^52–54

A separate study based on anatomic location and diagnosis with different margin sizes on the initial and subsequent stages would be required to see if the actual number of stages for complete tumor removed would be influenced by these factors. In our data set, and that of others, one could consider that the first stage on the ears and nose are potentially smaller to preserve tissue on sensitive anatomical areas as to minimize defects and facilitate a simpler repair; however, this may have resulted in more stages required to achieve tumor-free margins. Many reasons for fewer stages for SCC removal could be proposed, and this is also an area where further research is warranted. The difference is likely multifactorial, but knowing the exact margins on each stage, as mentioned above, and using that in the analysis may have changed the number of stages for tumor removal for various locations and SCC subtypes. The lips, hands/fingers, upper extremities, lower extremities, and trunk required fewer stages to clear tumor compared to other sites. Certainly, this is understandable for the extremities and the trunk, as larger initial margins can be taken on the first stage without compromising closure, except for possibly the pretibial area. The surgeon’s bias of knowing SCC can have dire consequences, especially on the lips and hand/fingers,^55–58 if not completely removed, may have encouraged larger margins on the initial and subsequent stages.^59–62 Alternatively, SCCs on the lips or hands/fingers may be detected and treated sooner and, therefore, may not have as many aggressive histopathological features as those on other body locations. Also, since there are more SCCs in smokers, and SCC overall required less stages to clear tumor in our data set, this may have resulted in fewer stages overall required for smokers than nonsmokers. Further study would be required to confirm this hypothesis.

Histological subtypes in MMS present unique difficulties. Many pathology reports of BCC listed more than one histological subtype. Bartos et al found that up to 35% of BCCs were of the “mixed” type.⁶³ Often, a different subtype of BCC may be found during MMS than what was reported on the original biopsy.^64–67 A recent report by Lim et al showed that many superficial BCCs required more stages because of this “histological drift.”⁶⁴ In our data set, the mixed groups, even aggressive mixed, were not associated with more stages than the mean for all BCCs. Further study of which anatomical sites and sizes have tumors with mixed subtypes and other aggressive subtypes could help in the anticipation of difficult tumors. SCC subtypes have not been well studied to see if there is “histological drift” during Mohs surgery and whether it affects the number of stages required for removal.

The main limitation of this study is its reliance on data from a single private practice with only two Mohs surgeons, which raises concerns about the generalizability of the findings beyond this specific setting. However, to address this limitation and enhance the validity of our conclusions, we compared our data to a larger dataset from Krishnan et al.¹³ This dataset consisted of a substantial sample size of 17,311 Mohs surgeries performed by 1845 surgeons between 2012 and 2014, providing a broader perspective. Notably, our dataset was statistically insignificant (P = 0.21) when compared to Krishnan et al, allowing us to utilize our mean number of stages taken (X = 1.77, N = 2788, SD = 0.95) for subgroup comparisons. Despite our efforts to improve the generalizability of our conclusions through this comparison, it should be acknowledged that the benchmark dataset used for comparison is almost a decade old. While it is unlikely that significant changes in surgeons’ practices have occurred during that time, it still raises questions about the current relevance and applicability of our findings.

In conclusion, this review shows that significant predictors exist for both increased and decreased numbers of MMS stages required to eradicate a tumor, which may help Mohs surgeons facilitate, plan, and allocate resources more effectively. Also, areas for further research in MMS are identified.

Disclosure statement/Funding

No funding or potential conflict of interest was reported by the authors.