Tumor Bed Margins Versus Specimen Margins in Oral Cavity Cancer: Too Close to Call?

Abstract

Introduction

The routine assessment of intraoperative margins has long been the standard of care for oral cavity cancers. However, there is a controversy surrounding the best method for sampling surgical margins. The aim of our study is to determine the precision of a new technique for sampling tumor bed margins (TBMs), to evaluate the impact on survival and the rate of free flap reconstructions.

Methods

This retrospective cohort study involved 156 patients with primary cancer of the tongue or floor of the mouth who underwent surgery as initial curative treatment. Patients were separated into 2 groups: one using an oriented TBM derived from Mohs’ technique, where the margins are taken from the tumor bed and identified with Vicryl sutures on both the specimen and the tumor bed, and the other using a specimen margins (SMs) driven technique, where the margins are taken from the specimen after the initial resection. Clinicopathologic features, including margin status, were compared for both groups and correlated with locoregional control. Precision of per-operative TBM sampling method was obtained.

Results

A total of 156 patients were included in the study, of which 80 were in TBM group and 76 were in SM group. Precision analysis showed that the oriented TBM technique pertained a 50% sensitivity, 96.6% specificity, 80% positive predictive value, and an 87.5% negative predictive value. Survival analysis revealed nonstatistically significant differences in both local control (86.88% vs 83.50%; P = .81) as well as local-regional control (82.57% vs 72.32%; P = .21). There was a significant difference in the rate of free flap-surgeries between the 2 groups (30% vs 64.5%; P < .001).

Conclusion

Our described oriented TBM technique has demonstrated reduced risk of free flap reconstructive surgery, increased precision, and similar prognostic in terms of local control, locoregional control, and disease-free survival when compared to the SM method.

Graphical abstract.

Introduction

Adequate surgical resection with negative margin is the standard of care for head and neck cancers, ¹ and failure to obtain negative surgical margins is a well-known risk factor for locoregional recurrences (LRRs).^2-4 Margins can be assessed on the specimen, called specimen margins (SMs) or on the tumor bed, called tumor bed margins (TBMs).^5-7 Positive margins are 1.7 times more likely to occur in squamous cell carcinoma (SCC) of the oral cavity when compared to other head and neck subsites. ⁸ Despite the importance of achieving negative peripheral and deep margins for the patient’s prognosis, there is controversy regarding the best method for margin samplings. The SM technique is currently the gold standard by the National Comprehensive Cancer Network (NCCN). ⁹ In this method, the resected specimen is inspected macroscopically by the pathologist and the surgeon. Tissue sampling is taken directly on the resected specimen for microscopic frozen section analysis in suspicious areas and additional TBM may be taken accordingly. Previous studies suggested that margins sampled from the specimen correlate better with prognosis as opposed to TBM.^10,11 Not only is there debate around the way a margin should be sampled, but there is also questionable reliability on the meaning of the peripheral margins. In fact, there is a poor correlation between margins sampled from the surgical bed and margins sampled directly on the specimen. Indeed, it was reported that, should a frozen section be positive, surgeons often inaccurately identify the corresponding site where the additional sampling has to be taken.^12,13 With oncologic and reconstructive goals in mind, the optimal margin sampling method should reduce the risk of LRR (appropriate oncologic resection), correctly identify the corresponding tumor bed should revision margins be necessary and limit the size of the surgical defects. Limited attention has been paid to optimize these factors. At our center, we use an adapted TBM technique, where peripheral TBM and deep TBM are assessed intraoperatively in an oriented fashion. The precision of our oriented TBM technique is yet to be determined and compared to the conventional SM method. The aims of this study are threefold: (1) to evaluate the precision of an oriented TBM technique for cancers of the oral tongue and/or floor of mouth (FOM; (2) to compare recurrences between the oriented TBM versus the SM; and (3) to compare the number of free flap surgeries required between the 2 techniques.

Method

Study Design and Population

This retrospective cohort study was conducted at 2 tertiary referral centers (Hôpital Enfant-Jésus and Hôtel Dieu de Québec, Université Laval, Quebec City, Canada) after approval from the Board of Ethics (#2020-5071). Medical and pathology records were reviewed for all patients treated surgically for primary SCC of the oral tongue and/or FOM from January 2015 to August 2020. Patients were excluded if they had the following: history of previous head and neck SCC, history of head and neck radiation, nonconventional variant of SCC (eg, verrucous carcinoma) neoadjuvant treatment, and no residual SCC in the final pathology specimen (eg, following excisional biopsy). The study cohort was divided into 2 groups based on the technique of margin assessment, the TBM (group 1) and the SM (group 2). Our referral centers have their specific ablative and reconstructive head and neck surgical oncology team, and each center use their respective method of margin assessment. Accordingly, group 1 and group 2 directly correspond to the hospital in which the patients were treated.

Margin Assessment Technique

In group 1, TBM sampling was performed in an oriented fashion. This method was derived from a Mohs’ technique that was used in our center from 2007 to 2012.^14-16 The tumor is delineated with an additional 10 mm of normal-appearing tissue using a monopolar cautery. With the specimen still attached to the deep tumor bed, the entire peripheral margins are sampled on the tumor bed (outside of the demarcated specimen) on 360°. Vicryl sutures are used to identify the exact limits (beginning and end) of each margin on both the specimen and the tumor bed. An additional deep margin from the entire tumor bed is resected and oriented. All margins are sent for frozen section examination. Re-resection is performed until negative peripheral and deep margins are confirmed by the pathologist (Figure 1). In this group, the final pathologic margin status (used for pathologic TNM staging) corresponds to the last frozen section analyses of the tumor bed. Postoperatively, all pathology reports including margin status are discussed and validated at the multidisciplinary head and neck tumor board.

Figure 1.

Oriented TBMs. (A) Marking of both the tumor bed and specimen after surgical delimitation. (B) Orientation of the specimen on a Telfa^® wound-dressing. (C) Marked surgical bed. TBM, tumor bed margin.

In group 2, margin sampling was performed using a SM technique. The specimen with 10 mm or more of normal-appearing tissue is resected with monopolar cautery and sent for macroscopic pathologic analysis. After resection of the surgical specimen, selected peripheral and deep TBMs were sampled, and sent for frozen section analysis. For the majority of cases, these analyses dictated the need for re-resections. Alternatively, intraoperative frozen analysis was done on the main specimen by the pathologist and guided revisions of the surgical bed. The final margin status was based on the final analysis of the resected specimen, regardless of the status of the frozen margins (Figure 2). This information was also revised and confirmed at the multidisciplinary head and neck tumor boards.

Figure 2.

SM technique. (A) Unmarked specimen after surgical delimitation. (B) Orientation of the specimen. (C) Unmarked surgical bed. SM, specimen margin.

Outcomes

Demographic data included patient’s age, sex, comorbidities, and alcohol and tobacco consumptions. Retrieved surgical and pathological variables included type of ablative and reconstructive surgery (including the need for a free flap, local flap, tracheostomy), clinical and pathological stages of the tumor (TNM), location of the primary tumor, presence or absence of lymphatic and vascular invasion (LVI), presence or absence of perineural invasion (PNI), pathologic grade of the tumor, tumor size, tumor thickness, depth of invasion, as well as presence or absence of adjuvant treatment (radiation with or without chemotherapy). The TNM staging was classified according to the seventh and eighth editions of the cancer staging manual from the American Joint Committee on Cancer (AJCC) depending on the year of surgery.

The precision of the TBM technique was defined by sensitivity, sensibility, as well as negative and positive predictive values (PPVs) compared to the resection margin status on the resected specimen. For all cases, the margin status was reported and analyzed as a binary variable (TBM and SM), where SM were considered as the true margins. SMs were classified as (1) positive margins (presence of cancerous cells at the edge of the margin), (2) very close margins (cancerous cells less than 1 mm from the edge), (3) close margins (cancerous cells located between 1 and 5 mm from the edge), or (4) negative margins (at least 5 mm of healthy tissue between the tumor and the edge). The presence of carcinoma in situ or high-grade dysplasia at the edge of the resected specimen was considered a positive margin. TBM were documented as a binary variable (positive or negative). Recurrences were categorized as local if they were at a contiguous site from initial cancer, regional if they were in the cervical lymph nodes or distant if they were outside of the head and neck region. All recurrences were measured from the date of the surgery to the date of the biopsy proving recurrence of SCC. Based on recurrences, secondary outcomes included local control, locoregional control, and disease-free survival (DFS).

Statistical Analysis

Demographic data, baseline characteristics of patients, and categorical variables were described as frequencies and percentages and compared using Chi square of Fisher’s exact tests as appropriate. Continuous variables were expressed as means and standard deviations or medians with interquartile range and compared with Student’s t test or Wilcoxon’s test as needed. The precision of each technique was calculated with the sensitivity, the specificity, the negative predictive values (NPVs), and the PPVs. Sensitivity analyses were performed to compare the TBM versus SM techniques. Sensitivity was also compared between the 2 surgeons from the team using the TBM method, to inquire if a difference in precision could be surgeon-dependent. The probability of survival without local recurrences (LRs), LRRs and DFS was estimated using the Kaplan-Meier method. Group differences were calculated using the log-rank test. Univariate cox proportional hazard regression models were performed to select potential confounding variables. These potential confounding variables were selected based on literature review as well as clinical relevance. Variables with a P value of less than .10 were included in the multivariate Cox regression model and used to evaluate all survival outcomes (LR, LRR, and DFS). Adjusted hazard ratios were calculated and reported with their 95% confidence intervals (CIs). Results with P values <.05 were considered statistically significant for the purpose of this study. All statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Study Population

A total of 156 patients who underwent surgery for a primary SCC of the oral tongue and/or FOM from January 2015 to August 2020 were included in the study cohort. Of these, 80 patients underwent the oriented TBM technique and were included in group 1 and 76 patients underwent the SM technique and were included in group 2. The median age of patients was 62.9 years and 61.3 years in groups 1 and 2, respectively (P = .48). The majority of patients included in the study had a T1 or T2 tumor (117/156 patients). The tumor size was slightly larger for patients in group 2 (group 1, 20.9 mm, vs group 2, 25.5 mm; P = .03). Other baseline characteristics were similar in both groups (Table 1). The median duration of follow-up for the patients alive at the last point of contact was 40.07 (IQR: 18.68-60.83) months for group 1 and 34.97 months (IQR: 15.58-50.50) for group 2.

Table 1.

Demographic Data.

Characteristics	TBM (N = 78)	SM (N = 76)	P value
Age	62.9 (12.8)	61.3 (14.7)	.48
Female	35	27	.24
Tumor stage (T)			.3
T1-T2	62	55
T3-T4	16	21
Smoking			.11
Never	23	14
Ever	17	13
Active	36	48
Alcohol			.19
Normal	48	37
Stopped	3	6
Active	25	32
Tumor location			.074
Tongue	48	55
Floor + tongue	23	11
Floor of mouth	7	10
Nodes			.49
N0	53	43
N1	10	13
N2+	15	19
PNI+	23	28	.35
LVI+	22	27	.35
Radiotherapy	36	39	.09
Chemotherapy	11	17	.18

Abbreviations: LVI, lymphatic and vascular invasion; PNI, perineural invasion; SM, specimen margin; TBM, tumor bed margin.

Margins

An average of 10.10 (95%: CI 8.89-11.32) margins per case were sampled for group 1 and 5.48 (95% CI: 4.95-6.01) margins per case for group 2 (P < .0001). TBMs were sampled in all patients included in group 1 and in 62/76 (81.52%) patients for group 2. When looking at the resected specimen regardless of tumor bed analyses, margins were positive in 17/76 (22.7%) patients for group 1 and in 12/75 (16%) patients for group 2 (P = .093). Final margin status was positive in 2/80 (2.5%) of patients for group 1 when taking the TBM revisions in consideration and in 12/75 (16%) for group 2 (P < .0001; Table 2). In group 1, the oriented technique pertained an overall sensitivity of 50% (95% CI: 24.65%-75.35%), a specificity of 96.61% (88.29%-99.59%), a NPV 87.69% (95% CI: 77.18%-94.53%) and a PPV of 80% (95% CI: 44.39-97.48%). This group was further analyzed according to the ablative surgeon. When surgeon number 1 was the ablative surgeon on the case (doing the tumor resection), the technique pertained a sensitivity of 66.7% (95% CI: 22.28%-95.67%) compared to 40% (95% CI: 12.16%-73.76%) when surgeon number 2 performed the resection. For surgeon number 2, sensitivity was 40% (95% CI: 12.16%-73.76%), specificity was 100% (95% CI: 92.60%-100%), NPV was 81.3% and PPV was 100% (95% CI: 47.82%-100%). In group 2, TBMs were sampled in 62 out of 76 patients. The SM technique pertained a sensitivity of 31.3%, a specificity of 100%, a NPV of 81.4%, and a PPV of 100%.

Table 2.

Margin Status.

Margins	Status	Group 1	Group 2	P value
Specimen margins	Negative	61.3% (46/75)	77.3% (58/75)	.0894
	Positive	22.7% (17/75)	16% (12/75)
	Close	16% (12/75)	6.7% (5/75)
Final margins	Negative	97.5% (78/80)	77.3% (58/75)	<.001
	Positive	2.5% (2/80)	16% (12/75)
	Close	0% (0/80)	6.7% (5/75)

Recurrences

In group 1, a total of 11/80 (13.75%) patients had a cancer recurrence (8 LRs, 3 regional recurrences, and no distant recurrences). In group 2, 19/76 (24.7%) patients had a recurrence (10 LRs, 8 regional recurrences, and 1 distant recurrence). In group 1, of the 17 patients with positive resected SMs, 3 had a LR (18%). In group 2, 4 out of the 12 positive SMs developed a LR (33%). No differences were found in terms of LR, LRR as well as DFS between the 2 groups (Figure 2). Results from the multivariate Cox regression model adjusted for adjuvant radiotherapy, pathological tumor staging, PNI, and LVI showed no statistically significant differences in terms of LR (HR: 1.305, 95% CI: 0.442-3.853), LRR (HR: 0.827, 95% CI: 0.367-1.866), and DFS (HR: 0.763, 95% CI: 0.345-1.689) between both groups.

Free Flap and Local Flap Surgeries

Patients in group 2 required a free flap reconstructive surgery more frequently than patients in group 1 (group 1, 24/80, vs group 2, 49/76 patients, P < .0001). In patients with early stage T1 and T2 tumor, 13/64 needed a free flap reconstruction in group 1 versus 31/55 in group 2 (P < .0001) (Figure 3). Fourteen out of 80 patients (17.5%) in group 1 had a local flap compared to 6 out of 76 patients (7.89%) in group 2 (P = .0729).

Figure 3.

Kaplan-Meier survival curves for local recurrences (A), locoregional recurrences (B), and disease-free survival (C).

Tracheostomy

In group 1, 27/80 (33.75%) patients required a tracheostomy compared to 46/76 (60.53%) patients in group 2 (P = .0008).

Discussion

Precision of Margin Sampling

Surgical resection with clear margins is the standard of care for oral cavity cancers. Obtaining an initial negative margin is the best case scenario in terms of prognostic for head and neck cancer patients. For multiple reasons, the intricacies of head and neck surgical oncology do not always allow surgeons to achieve negative margins on the first attempt. Some authors have implied that re-resections have little benefit in improving LRs. Patients with initially positive margins, in whom additional margins are taken, will not have as good a prognosis as patients with initially negative margins. However, this does not translate as a complete absence of benefit.^10,12,17 In a cohort study including 344 patients with oral cancers, Buchakjian et al reported that the risk of LRs was the lowest when SMs were negative, higher when positive margins were revised to negative, and the highest when final margins were not cleared. ¹⁸ For these reasons, the need for a reliable and reproductible technique for margin sampling is advisable. Our study compared the precision of an oriented TBM technique to a SM technique. The use of an oriented TBM technique pertained a better sensitivity than the SM method, with a sensitivity of 50% and 30%, respectively. The performance of this method improved when the surgery was performed by the surgeon who developed the technique (surgeon number 1). This implies a certain operator dependent variation in the application of the technique, due to consistency and experience. These results compare favorably to previously published studies, all reporting lower sensitivities, ranging from 15% to 32%.^12,13 The poor performance of TBM has been highlighted on many occasions in the literature. In their multicenter retrospective cohort study, Maxwell et al reported that 95 patients who had TBMs only had a sensitivity of 24.2% (95% CI: 16%-34%) and a specificity of 92% (95% CI: 85%-97%) when SMs were considered the true margin. ¹² Similarly, Buchakjian et al found that TBM were 35% sensitive in identifying positive SMs. ¹⁸ Another study published by Chang et al showed that TBM failed to identify positive glossectomy margins in 67% of patients. ¹⁰ In fact, the lack of accurate localization of positive margins were reported by Kerawala and Ong. This group specifically performed a study on TBM relocation. Surgeons were asked to indicate the sites of initial sampling of TBMs and to re-identify each site 5 minutes later. Interestingly, an error of more than 1 cm was found in 32% of cases. ¹⁹ As the standard technique for sampling TBMs revealed to be unreliable, the necessity of alternative techniques is undeniable. Some authors have described the placement of surgical clips on the surgical bed,^20,21 others, systematic cavity shavings.^22,23 At our center, we performed an oriented TBM technique. van Lanschot et al ²⁴ have also described the feasibility of a similar technique using paired tags on both specimen and surgical beds. In their study, the main specimen was analyzed by both the pathologist and the surgeon intraoperatively. Additional resections were performed based on these markers. After those additional resections, the status of resection margins was improved in 28 out of 31 cases and surgeons also reported an easier relocalization of the positive resection margin from the specimen to the wound bed. ²⁴ This technique is similar to the one described herein, with the difference that we used sutures, a readily available material, to mark both the specimen and surgical bed as opposed to surgical clips.

In addition to significantly increase precision with the oriented TBM technique, our study showed that the specimen-driven approach was correlated with a significantly increased risk of needing a free flap reconstruction, both for early stage (T1-T2) and advanced stage (T3-T4) oral tongue and FOM cancers. One can argue that tumors were slightly smaller in group 1 (20.9 mm vs 25.5 mm for group 2; P = .03), but the difference in tumor size alone cannot explain the discrepancy in the need of free flap reconstruction. Only 24/80 patients required a free flap reconstruction in the TBM group compared to 49/76 patients in the SM group. Moreover, it is well known that patients with involvement of the FOM are more at risk of requiring a free flap reconstruction to achieve a water-tight closure and avoid salivary fistula, especially when a neck dissection is concomitantly performed. In our cohort, 30/80 patients in the oriented TBM group had a resection of the FOM, with or without glossectomy, compared to 22/76 patients in the SM group. Yet, the number of free flap reconstructions was significantly lower in the TBM group. Furthermore, the higher incidence of local flap reconstructions in the TBM cohort (group 1), as opposed to the SM cohort (group 2), suggests that the surgical deficits were smaller in the former. In addition, the increased frequency of tracheostomies as well as the increase need for free flap reconstruction in group 2 could imply that more extensive resections were performed in this group, potentially indicating larger surgical deficits. Considering that the initial negative SM confers the best prognosis and assuming that additional resections are not beneficial, the tendency may be to take wider margins to increase the chances of achieving negative SMs, but as a downside, creating larger oral cavity defect. Not only do these major surgeries have a significant impact on patients’ speech and swallowing functions and quality of life,^25,26 but they also increase the burden on the healthcare system with longer hospital length of stays as well as higher rates of complications and postoperative readmissions. ²⁷ Head and neck cancer resections are particularly challenging: oncologic resections must be large enough to obtain an optimal prognosis and limited enough to preserve as much function as possible. Surgeons should have to ponder the real benefit from taking bigger SMs to feel reassured of achieving a margin negativity at the cost of patients’ speech and swallowing abilities, quality of life, and increased morbidity. Even with wider initial margins, there is no guarantee that microscopic evaluation of margins will be negative, and this oriented TBM method emphasizes the importance of a standardized approach to obtain margin control.

In previous studies, the use of intraoperative margins has been associated with increased operative times and costs. ²⁸ Moreover, these additional costs can be carried outside the operating room, because false negatives result in subsequent escalation of therapy. ²⁹ In addition, TBMs have been associated almost systematically with a worse prognosis, and SMs have been correlated with LRs whereas TBMs have not.^10,12,30-33 In contrast to these studies, we found that the use of oriented TBM was not associated with a worse prognosis for local control, locoregional control as well as DFS when compared to SM. Improvement in prognosis with additional per-operative margin resections has been reported in other studies as well.^34,35 Several theories may explain the variability of these results. First, in previous studies, the revised margins may not adequately clear the residual disease. Indeed, considering the poor sensitivity of the standard tumor bed sampling technique, one can hardly assume that an adequate revision is accomplished, and that the re-resection was sampled in the correct area. In our case, adding the orientation of the margins on both the specimen and the tumor bed pertains a higher sensibility and precision. This can prevent errors caused by both the surgeon’s memory bias as well as the tissue contraction in the complex 3-dimensional anatomy of the head and neck. Second, the meaning of a negative TBM is yet to be understood. ¹⁷ In fact, a TBM is analyzed in a binary fashion, positive or negative, as opposed to the SM, who can be classified as positive, very close, close, or negative. Distance cannot be measured on a frozen section taken from the tumor bed. Without the ability to measure the distance to the edge, like one would classically do on a specimen, it is unclear how a frozen margin should be interpreted. When a SM is negative and the additional TBM is also negative, it is without argument a negative margin. But how should we classify the other scenarios? For example, if a specimen has a close margin (between 2 and 5 mm) with negative TBM, would this be considered a negative margin? If a TBM turn out negative after multiple revisions, should that margin also be classified as a negative margin or a close one? The latter situation can lead to subsequent escalation of therapy. There is currently no consensus on how to categorize these patients and adjuvant treatment is center dependent. As described, in our study, a negative TBM was considered a negative margin, regardless of the status of the SM. Even without a measurable margin clearance of 5 mm, our TBM cohort did not have worse prognosis than the SM group, and escalation of therapy was avoided. Last, to explain the disparity in the literature regarding prognosis, we believe that a positive margin may be a surrogate for a biologically high-risk tumor. ³⁶ PNI, LVI, a high worse pattern of invasion grade (WPOI-5) as well as higher grade of tumors are pathologic adverse features that have been associated with increased rates of positive margins.^2,37,38 Other immune and genetic alterations could also be linked to both positive margins and recurrences, but this remains to be studied more thoroughly. ³⁹ Understating that a microscopic positive margin on formalin-fixed tissue and tumor genetic and biology may be risk factors that cannot be completely eradicated even by adequate intraoperative revision of the surgical bed, the real question that remains is whether revised margins have a better prognosis than unrevised positive margins. There is currently lack of sufficient power in our study and previously published articles to answer this question. Moreover, to answer this question adequately, we must ensure that revisions of margins on the tumor bed are performed as accurately as possible to eliminate potential confounding factors. While relying exclusively on tumor-bed margins may not be optimal, we believe that a combination of both specimen and tumor-bed per-operative circumferential margin status and distance to margin should be adopted to provide the surgeon with the best guidance since visually (macroscopically) unsuspected positive margins are inevitable in some patients. The evidence is still too scarce to adopt a “no-revision, rely-on-radiotherapy” approach, and the focus of future studies should be on improving the sampling technique as well as obtaining better quality data and adjusted survival analyses from larger cohorts of patients.

Limitations

Our study has some limitations. First, some potential confounders such as margin distances were not systematically reported in each pathology reports as pathologists only described distances when inferior to 5 mm. The WPOI was also absent in many cases. Second, because of the retrospective nature of the study, we cannot evaluate the exactness and accuracy with which the oriented tumor bed technique was performed in each case. A prospective study with standardization of the technique would allow for better reproducibility, and therefore, could increase the sensitivity of the herein described oriented TBM technique. Last, the sampling method for TBMs in group 2 was far from standardized. On some occasions, revisions were driven by analysis of tumor bed frozen section margins whereas they were driven by analysis of the main specimen by the pathologist on other occasions. This variation in practice may have been caused by the increasing literature favoring analysis of SMs.^8,12,30 The observed variability in the margin sampling technique within group 2 not only represents a limitation of our study but also likely mirrors the broader lack of standardization in current surgical practices. Frequently, surgeons in group 2 did not sample and evaluate peripheral margins, a practice that may result in disproportionately large resections. Such extensive surgeries, while more aggressive, do not necessarily correlate with improved prognostic outcomes. This raises concerns about the efficacy and appropriateness of these techniques in achieving the desired clinical objectives.

Conclusion

In this study, we propose an oriented TBM technique to improve peripheral and deep margins sampling for intraoperative frozen section analyses. While many studies point out that negative SMs are correlated with a better prognosis, there is still a need to adequately identify residual cancer on the tumor bed as positive margins are inevitable in some cases. Our described oriented TBM technique has demonstrated reduced risk of free flap reconstructive surgery, increased precision, and similar prognostic in terms of local control, locoregional control, and disease-free survival when compared to the SM method. While this technique has to be further studied, it is a step toward improving current margins sampling and assessment techniques.