The utility of intra-oral ultrasound in improving deep margin clearance of oral tongue cancer resections

Abstract

Objectives:

To investigate the potential utility of intra-oral ultrasound (IOUS) in guiding deep margin clearance and measuring depth of invasion (DOI) of oral tongue carcinomas (OTC).

Materials and Methods:

Retrospective chart review of consecutive patients with T1-T3 OTC who underwent intraoperative ultrasound-guided resection and a comparator group that had undergone resection without the use of IOUS both by a single surgeon. Data was extracted from operative, pathology and radiology reports. Deep margins and DOI were reviewed by a dedicated head and neck pathologist. Correlation between histologic and ultrasound DOI was assessed using Pearson correlation.

Results:

A total of 23 patients were included in the study cohort with a comparator group of 21 patients in the control group. None of the patients in the study cohort had a positive (cut-through) deep margin and the mean deep margin clearance was 8.5 ± 4.9 and 6.7 ± 3.8 for the IOUS and non-IOUS groups respectively (p-value 0.18) showing a non-significant improvement in the IOUS group. As a secondary outcome, there was a strong correlation between histologic and ultrasound DOI (0.9449).

Conclusion:

Ultrasound appears to be a potentially effective tool in guiding OTC resections. In this small series, IOUS facilitated deep margin clearance and resulted in a non-statistically significant increase in deep margin clearance. Intraoral ultrasound can accurately measure lesional DOI.

Introduction

Oral tongue cancer (OTC) remains a significant health issue in the United States and around the world [1]. The mainstay of treatment is complete surgical resection of the primary site with adequate margins and neck dissection based upon clinical nodal disease status and pathologic risk factors for occult nodal disease. Although there is controversy regarding what radial distance constitutes an adequate margin in oral cancer resection, the majority of surgeons support a radial distance of 5 mm on final pathologic examination in a published survey we recently completed [2].

Obtaining adequate margins in oral tongue resections remains a challenge as oral cavity cancer resections have been shown to have a higher rate of positive surgical margins than all other solid tumors except for ovarian and prostate (3rd highest with a prevalence of 12.75%) [3]. For OTC, it is the deep resection margin that poses the greatest challenge. Indeed, a retrospective series of 301 oral cavity and oropharyngeal resections noted positive margins 23% of the time (<1mm; with 13% noted to have histological cut through) and the deep margin was noted to be the source of the positive margin 87% of the time [4].

The use of intraoperative frozen section (FS) margin analysis is accepted as the standard of care for oral cancer resections, although conclusive evidence of its benefit has yet to be proven [5]. A recent meta-analysis has shown that an initially positive margin converted to negative under FS guidance (R1 to R0) is not equivalent to an initially negative (R0) resection and has a 2.5 times worse 5-year local recurrence free survival (LRFS) [6]. Thus, obtaining a negative margin (>5 mm) on the “first pass,” initial resection is critical; yet, there are no widely used, reliable tools to augment the surgeon’s ability to accomplish this goal.

Traditional techniques, in addition to FS, that have been used to obtain clear margins include review of preoperative imaging and intraoperative visual evaluation and manual palpation with a proposed minimum of 1 cm radial gross margin, where possible. However, these techniques have limitations, namely, the subjectivity and imprecision inherent with inspection and palpation. Furthermore, pre-operative imaging (MRI and CT) may not show OTC very well, especially when tumors are < 5 mm [7]. Newer techniques of margin evaluation currently under investigation include next-generation sequencing-based biomarkers [8,9], optical imaging [10], and intra-oral ultrasound (IOUS) [11–14]. Recent research on optical imaging has shown theoretical usefulness with surface mucosal margins as evidenced in early clinical trials [10], however, these approaches are limited in their ability to assess normal tissue beyond the tumor-normal mucosal interface and are unable to visualize the deepest point of tumor infiltration while the tumor resides in situ. By contrast, several studies, including those from our institution, have shown the feasibility and early outcomes of using IOUS to provide additional information in visualizing the deep tumor-normal tissue interface prior to starting resection, along with methods for determining tumor depth of invasion (DOI)/thickness [11–14]. Intraoperative ultrasound is an evolving modality with respect to oral cancer resections that offers potential advantages, including 1) allowing the surgeon to see then entire lesion, not just limited to the mucosal surface, but including the tumor’s deepest extent, 2) accurately measures the depth of invasion and the deepest extent of the tumor, 3) provides the ability to evaluate the deep margin at any point during the procedure, and if inadequate, allow real-time adjustments before the specimen is resected, 4) avoid the need for preoperative injection, and 5) is widely available and requires no additional capital expense to most institutions.

In this study, we extend on our initial feasibility study utilizing IOUS. Here, we report a series of OTC resections with and without the use of IOUS as a single surgeon experience with a primary endpoint showing an improved deep margin clearance and compare the results to deep margin clearance in a comparator group of patients of the senior author. A secondary endpoint includes the correlation of IOUS measurement of DOI to histological measurement.

Materials and methods

Patient selection

This is a case series study. Patients who presented to the senior author with T1-T3 OTC underwent IOUS-guided resection of OTC were identified and included in the study group. All patients were treated by the same surgeon. Pathology slides were re-reviewed by a single pathologist, one at each of the 2 participating institutions, blinded to the previously reported deep margin and DOI on pathology and radiology reports. IRB approval was obtained for retrospective data acquisition. Approval for the use of IOUS was waived by the Massachusetts Eye and Ear IRB as it was felt to represent surgical innovation and not a clinical trial given that ultrasound is already used in the management of head and neck cancers.

A comparator group was established by retrospectively collecting data from a matched cohort of T1-T3 OTC patients who underwent resection without IOUS from the senior author’s prior institution (Saint Louis University). These patients were pulled from a previous database that was available for the Saint Louis University contributors. IRB approval was obtained from The Saint Louis University IRB. All lesions in both groups had not been treated previously.

Surgical procedure

The surgical approach was transoral in all but one case in the IOUS series. At the time of surgery, the tumors were visualized, palpated, and a gross margin of 1.0–1.5 cm was outlined circumferentially. A broadband compact linear array ultrasound transducer (L15–7io; Philips) enclosed in a sterile plastic cover was introduced into the surgical field, placed on the surface of the target lesion and used to evaluate the 3-dimensional extent of the tumor. Figure 1 illustrates the intraoperative image of a case in this series. In particular, the tumor thickness, DOI and deep tumor extent of the tumor in Figure 1 were evaluated and captured (Figure 2). This information augmented the visual and tactile evaluation of the tumor needed to plan the deep extent of the resection but did not replace it. The resection then proceeded in a typical anterior to posterior fashion using monopolar electrocautery. At the midpoint of the resection the ultrasound probe was brought back into the surgical field to estimate the distance of the deep resection margin across the entire deep surface of the resection to that point (Figure 3). This margin was also assessed at this point of the resection using palpation. The resection was then completed in standard fashion. Once the resection was completed the specimen was re-examined ex vivo using the ultrasound probe, once again assessing the deep margin of resection across the entire deep surface (Figure 4). The estimated overall additional time to the procedure with the use of IOUS was 5–10 min. The resection specimen was then taken by the surgeon to the FS lab and evaluated with the attending pathologist. Gross evaluation was performed together, and consensus was developed on where to take radial FS from the main specimen (specimen-based margins). The goal in all resections was a 5 mm or greater microscopic surgical resection margin across the entirety of the specimen. The specimen was then sent for permanent fixed evaluation and final margin status was determined as per standard practice.

Figure 1.

Exophytic lesion of the right lateral oral tongue.

Figure 2.

Exophytic tumor completely visualized by intraoperative ultrasound just prior to resection. Horizontal line represents hypothetical line from surrounding normal mucosa. Dashed line represents TT (7 mm on US; 6 mm on final pathology). Solid line represents DOI (2.79 mm on US; 2 mm on final pathology). [Reprinted from Oral Oncology, Vol 99, Bulbul et al, Understanding approaches to measurement and impact of depth of invasion of oral cavity cancers: A survey of American Head and Neck Society Membership, Figure 1, Copyright (2019), with permission from Elsevier]

Figure 3.

Deep margin halfway through resection. The echogenic line represents air within the plane of resection (long arrow on bottom left corner). The thin blue line connecting the electronic calipers represents deep margin clearance (10 mm). Short arrow marks the end of the midway point of resection. Star represents the tumor.

Figure 4.

Deep margin post-resection. Echogenic line marks the edge of the specimen (long solid arrow). Short solid arrow marks the peripheral mucosal tumor margin. Thin arrows mark the deep edge of tumor. Star marks the exophytic tumor overlapping the mucosal surface.

The performance of concurrent neck dissection was dependent on several factors including tumor grade, the known presence of perineural invasion or lymphovascular invasion preoperatively, a DOI greater than or equal to 4 mm, and the presence of clinical or radiographic nodal disease. Flap reconstruction was performed depending on the opinion of the surgical team based upon the dimension and location of the defect. Adjuvant therapy was employed according to NCCN guidelines and where pathologic criteria suggested increased risk of local or regional recurrence.

Statistical analysis

Categorical variables were compared using chi-squared analysis whereas continuous variables were compared using t-test. Correlation of IOUS to histologic DOI was assessed using Pearson correlation. Stata13 was used for all statistical analyses (StataCorp LP, College Station, Texas, USA). R 4.0.2 was used for graphics (The R Foundation for Statistical Computing Platform).

Results

A total of 23 patients that underwent resection using IOUS and 21 comparators that were operated without IOUS were included in our study (Table 1). The mean closest margins were 6.3 ± 2.8 and 4.3 ± 2.7 for the IOUS and non-IOUS groups (p-value for the difference, 0.018), respectively. The mean deep margins were 8.5 ± 4.9 mm and 6.7 ± 3.8 for the IOUS and non-IOUS groups (p-value for the difference, 0.18), respectively. Five of 23 (22%) of the IOUS patients had a deep margin<5 mm; one at 1 mm, one at 3 mm, and three at 4 mm versus 7 of the 21 (33%) patients in the non-IOUS group (p-value 0.39) (Figure 5). With respect to the IOUS group, the single 1 mm margin was a peripheral mucosal margin and deep margin that was reported on final pathology and not on frozen section analysis. The one 3 mm and three 4 mm margins were noted on frozen section and a decision was made intraoperatively not to resect these patients further as the margins were felt adequate given the lack of clearly proven benefit of additional resection. There were no cases of microscopic tumor cut-through, i.e. positive margins (risk: 0%; 95% CI: 0 – 13%) in the IOUS group and one case of cut through in the control group.

Table 1.

Baseline Characteristics of the US and non-US group. US = Ultrasound, DOI = Depth of Invasion.

Characteristic	US-guided Resection (N = 23)	No Ultrasound (N = 21)	p-value
Age, mean ± SD	59.1 ± 17.2	63.9 ± 12.2	0.29
Gender
Female	8 (35%)	11 (52%)	0.24
Male	15 (65%)	10 (48%)
T-stage
T1	13 (57%)	8 (38%)	0.40
T2	8 (35%)	9 (43%)
T3	2 (9%)	4 (19%)
N-stage
N0	20 (87%)	15 (71%)	0.047
N+	3 (13%)	6 (29%)
Closest Margin (<5, >5)
>5 mm	16 (70%)	10 (48%)	0.14
<5 mm	7 (30%)	11 (52%)
Closest Margin (mm), mean ± SD	6.3 ± 2.8	4.3 ± 2.7	0.018
LVI
No	22 (96%)	20 (95%)	0.95
Yes	1 (4%)	1 (5%)
PNI
No	14 (61%)	11 (52%)	0.57
Yes	9 (39%)	10 (48%)	Yes
Deep Margin (<5, >5)
>5 mm	18 (78%)	14 (67%)	0.39
<5 mm	5 (22%)	7 (33%)
Deep Margin (mm), mean ± SD	8.5 ± 4.9	6.7 ± 3.8	0.18
Histologic DOI (mm), mean ± SD	6.4 ± 4.3	10.8 ± 6.2	0.009
US DOI (mm), mean ± SD	6.8 ± 3.3	NA
Adjuvant
No	15 (68%)	10 (50%)	0.013
Yes	7 (32%)	10 (50%)

Figure 5.

Boxplot showing the distribution of the deep margins by group.The small dots are individual data points (patients) while the big dots represent aggregate data points.

The mean DOI in the IOUS group was 6.4 ± 4.3 mm and in the non-IOUS group 10.8 ± 6.2 (p = 0.009). In addition, the non-IOUS had a greater number of T2–3 patients than the IOUS group. There was a strong correlation between DOI as measured by IOUS and histologic analysis (Pearson correlation coefficient = 0.9449, p < 0.0001).

Discussion

In this single-surgeon series evaluating the utility of IOUS in OTC resections, we found promising results indicating improved ability to clear the deep margin of resection, as evidenced by: 1) the mean overall and deep margin clearance; 2) rate of negative (≥5 mm) deep margin (78%); and 3) absence of frankly positive deep margins. Although not statistically significant, this series did show a trend towards improvement in deep margin clearance and rate of margins<5 mm over an imperfectly matched cohort from the senior surgeon’s series without the use of IOUS. However, compared to historically reported rates of positive margins based on published retrospective studies [4], the IOUS cohort also demonstrated an apparent improvement in margin clearance, especially the deep margin, over the standard-of-care approach. In addition, measurement of the tumor DOI using IOUS was found to significantly correlate with DOI as definitively measured on final histological analysis of the tumor. The significance of this is two-fold: 1) DOI has now become part of the new clinical staging criteria (AJCC 8th edition) and thus ultrasound may be a useful imaging modality for accurate pre-operative staging/work-up; and 2) DOI is widely used by surgeons to inform the decision to perform neck dissection [15] and IOUS may become of wide use for this purpose as more studies show promising results. Exactly what depth of invasion is the threshold for performing elective neck dissection is beyond the scope of this report. However, we felt it is important to highlight the strong correlation of IOUS to DOI on final histologic evaluation, as this finding 1) broadly supports the fidelity of IOUS in capturing/representing deep margin histopathologic features, and 2) is potentially important in the discussions of the relationship of DOI to elective neck dissection. In the future, the use of IOUS could spare the patients the need for separate subsequent surgical intervention for neck dissection until final pathology on the primary tumor specimen is available on which to determine depth of invasion. Notably, the advantage of performing intraoperative ultrasound assessment of oral tongue DOI in the operating room is largely based on patient comfort, as in many cases, the lesion is too painful to allow US evaluation without some type of anesthetic.

Previous investigators have shown some promise with IOUS in clearing the deep margin in OTC resections. Helbig and coworkers described a technique where IOUS was used to detect a braided surgical suture placed 1–2 mm away from the deep margin of the tumor [13]. Baek [11] and Kodama [12] both employed the use of ultrasound to guide the placement of hollow needles driven through the tumor to a deep distance of 1.5 and 1.0 cm respectively, using the deep tip of the needles as the means of assessing the location of the deep margin cuts. The study by Baek demonstrated a statistically significant difference in radial deep margin clearance when compared to a stage-matched control cohort [11]. Likewise, Kodama showed in their small series that a deep margin clearance of>10 mm was reliably achieved with a similar approach [12]. Our group has previously reported a feasibility study of 12 patients where IOUS was employed and all but one achieved a deep margin>5 mm, with one patient with a deep margin clearance of 4 mm [14]. Unlike the other 3 previously reported studies, this study and our prior series did not require the placement of deep suture around or needles through the tumor, techniques that we felt may be inconsistently implemented, potentially dangerous to the surgeon (from unintended needle injury), and incompatible with a complete oncologic resection. In addition, ultrasound technology continues to evolve and the recent deployment of an 18 MHz Philips probe has resulted in images of superior resolution, as demonstrated in Figure 6, which illustrates the use of this probe in a recent case not part of this series.

Figure 6.

A. Initial survey of T2 oral tongue cancer with an 18 MHz ultrasound probe. The irregularly shaped hypoechoic soft tissue represents the tumor, with tumor thickness demarcated by the plus-shaped electronic calipers. B. Midway point of resection. The X-shaped calipers demarcate the deep margin clearance. C. Evaluation of completed resection.

In spite of the widespread use of FS in oral cancer resections [2], there continues to be little evidence as to the efficacy of this modality [5]. Although the accuracy of analysis at FS is very high, the use of FS to revise intraoperative margins has yet to show a clear oncologic advantage. However, a majority of surgeons surveyed answered “yes” when asked if a revised initially-positive surgical margin “to clear” is equal to the oncologic safety of an initially clear margin [2]. The inability of FS driven margin revision to result in a LRFS equal to an initially negative resection margin is probably related to issues with technique (patient-based versus specimen-based) and the inherent possibility of sampling error regardless of which of the two approaches above are employed [16]. Recent reports in the literature have pointed to the importance of the analysis of the margin status of the main surgical resection specimen as the best predictor of local disease control [17]. Given the observation, it is critical that the surgeon clear the tumor with an adequate margin during the initial attempt at resection and that the main surgical specimen be used to determine margin status, both at the time of surgery and on final pathologic analysis. In this series intraoperative ultrasound facilitated clearance of the deep margin and avoided of the need to revise that margin.

We acknowledge that there are several limitations of this study. First, this study lacks an ideally matched comparison group as our comparator has a greater proportion of T2–3 patients and a statistically significant greater DOI. We might be able to make stronger inferences if a more closely matched comparison group had been available, but it was not at this time. In spite of this limitation, the finding of nearly 2 mm greater mean deep margin clearance in the IOUS group compared to the non-IOUS group, the absence of tumor cut through, and the low proportion of patients with a close (<5mm) margin is noteworthy and displays promise. Given that the goal of oral cancer resections is not the widest margin possible but rather the approach that results in the greater number of patients with adequate deep margin clearance while sparing normal tissue to preserve function, this increase may be important, especially when coupled with the observed > 10% improvement in patients with < 5 mm deep margin in the IOUS group over the comparator group. To use a primary outcome measure of the average deep radial margin distance may not be the best metric, but rather what may be a better goal is the percent of patients within the range of 5–10 mm for the radial deep margin. Our other comparator group is the previously published, high rate of positive margins overall in oral cancer resections [3] and the high frequency of positive deep margins seen in oral cavity cancer resections [4]. Future studies must include such a well-matched comparator cohort in order to fortify and validate the hypotheses suggested by this study. It is in this specific vein of having a precise method to estimate margin that is “unseen” where IOUS offers potential benefit to even the most experienced surgeons by making the resection margins a more objective calculation rather than a subjective estimate based on palpation. The demonstration of such an advantage has yet to be proven and should be the focus of well-designed future studies. In spite of these weaknesses, we feel that the non-invasive approach of IOUS, which adds minimal time to the overall time of the procedure, should be more widely investigated to determine if indeed there is an advantage to its use. If found effective, this could be a simple approach that could result in fewer patients with positive margins, a decreased need for adjuvant therapy, with decreased incidence of excessive tissue resection and ultimately, better functional and oncologic outcomes for our patients. It should be noted that patients with T4 disease were not included in this series for several reasons. First, tumor extending into the mandible would be inaccessible for evaluation with this technique. In addition, such patients typically undergo en block resection (possibly in continuity with neck dissection), prohibiting deep margin evaluation intraoperatively. Finally, patients with advanced disease will most certainly require adjuvant therapy – one potential advantage of the ultrasound modality is avoiding a positive margin in patients with early and intermediate staged disease who could potentially otherwise be spared adjuvant therapy.

Conclusion

IOUS appears to be a potential tool in guiding OTC resections based on the results of this series, where it accurately measured DOI and improved deep margin clearance and tumor cut through rates over those seen in our control group and that reported in the literature. As more interest in this approach is generated, studies including a prospective randomized controlled study would be required to prove that its efficacy clearly improves disease control and survival while maximally preserving function.