- Email: [email protected]
Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma
Oral Oncology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology
Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma Gregor Heiduschka a,b,⇑, Sohaib A. Virk c, Carsten E. Palme a, Sydney Ch’ng a,g, Michael Elliot a, Ruta Gupta d, Jonathan Clark a,e,f a
Department of Head and Neck Surgery, Chris O’Brien Lifehouse, Camperdown, New South Wales, Australia Department of Otorhinolaryngology – Head and Neck Surgery, Medical University of Vienna, Austria University of New South Wales, Randwick, New South Wales, Australia d Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia e Central Clinical School, The University of Sydney, Sydney, New South Wales, Australia f South Western Sydney Clinical School, University of New South Wales, Sydney, New South Wales, Australia g The Institute of Academic Surgery at RPA, University of Sydney, NSW, Australia b c
a r t i c l e
i n f o
Article history: Received 8 September 2015 Received in revised form 9 January 2016 Accepted 13 January 2016 Available online xxxx Keywords: OSCC Margin Thickness Prognostic marker
s u m m a r y Objectives: To assess whether small oral squamous cell carcinomas (OSCC) require the same margin clearance as large tumors. We evaluated the association between the ratio of the closest margin to tumor size (MSR) and tumor thickness (MTR) with local control and survival. Methods and methods: The clinicopathologic and follow up data were obtained for 501 OSCC patients who had surgical resection with curative intent at our institution. MTR and MSR were computed and their associations with local control and survival were assessed using multivariable Cox-regression model. Survival curves were generated using the Kaplan–Meier method. Results: MTR was a better predictor of disease control than MSR. MTR was a predictor of local failure (p = 0.033) and disease specific death (p = 0.038) after adjusting for perineural invasion, lymphovascular involvement, nodal status, and radiotherapy. A threshold MTR value of 0.3 was identified, above which the risk of local recurrence was low. Conclusion: The ratio of margin to tumor thickness was an independent predictor for local recurrence and disease specific death in this cohort. A MTR > 0.3 can serve as a useful tool for adjuvant therapy planning as it combines tumor thickness and margin clearance, two well established prognostic factors. The minimum safe margin can be calculated by multiplying the tumor thickness by 0.3. Further prospective studies in other institutions are warranted to confirm the prognostic utility of MTR and assess the generalizability of our threshold values. Ó 2016 Elsevier Ltd. All rights reserved.
Introduction Surgical resection is the mainstay of treatment for oral squamous cell carcinoma (OSCC). Satisfactory oncologic resections require negative and adequate resection margins. A microscopic margin of 5 mm has traditionally been accepted as adequate for various oral cavity sub-sites [1,2]. This definition, however, of adequate surgical resection remains controversial as some authors propose 2 mm to be sufficient [3] while others advocate for more than 7 mm [4]. Ch’ng et al. [5] and Brandwein-Gensler et al. [6] ⇑ Corresponding author at: Sydney Head & Neck Cancer Institute, The Chris O’Brien Lifehouse, Royal Prince Alfred Hospital, Missenden Road, Camperdown, NSW 2050, Australia. Tel.: +61 8514 0109; fax: +61 9519 9214. E-mail address: [email protected] (G. Heiduschka).
have suggested that the adequacy of surgical resections need to be considered in the context of various pathological characteristics such as differentiation, growth pattern, and perineural invasion (PNI). It is intuitive that larger tumors with an infiltrative growth pattern would require larger margins compared to smaller tumors, particularly those with pushing margins. Thus a ratio that takes into account well established prognostic factors such as tumor size, tumor thickness, and distance to the closest margin may help evaluate whether a certain margin can be considered as safe. The clinical utility of ratios of two related pathological variables is becoming increasingly recognized in oncology. An important example of this is lymph node ratio (LNR) which is a strong predictor of survival in oral cancer [7–9]. LNR incorporates not only the number of lymph node metastases, but also the total number of nodes counted by the pathologist, which enables information on
http://dx.doi.org/10.1016/j.oraloncology.2016.01.010 1368-8375/Ó 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
2
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
burden of disease, comprehensiveness of surgical resection, and thoroughness of pathological examination to be combined into a single parameter. Similar guidelines regarding adequate surgical resections are well established in melanoma based on depth of invasion [10]. Although tumor thickness and depth of invasion have been shown to be important prognostic factors in OSCC, the relation between tumor size or thickness and the resection margins has been not been studied [11–14]. The aim of this study is to evaluate whether a ratio between the resection margin and tumor thickness or tumor size is a predictor of local control or survival in OSCC and whether a clinically useful threshold value that may aid adjuvant therapy planning can be identified. Methods Study population The Sydney Head and Neck Cancer Institute has maintained a prospective database including the clinical, pathologic and follow-up data of all patients treated in the Department of Head and Neck Surgery, Royal Prince Alfred Hospital and Chris O’Brien Lifehouse since October 1987. After obtaining institutional ethics approval, clinico-pathological data for all patients with OSCC treated between October 1987 and December 2014 was extracted from the database. Histologic variables Tumor size was defined as the largest dimension of the tumor as measured in millimeters during macroscopic examination and confirmed microscopically for tumors less than 10 mm in maximum dimension. Tumor thickness was measured on formalin fixed paraffin embedded sections stained with haematoxylin and eosin to the nearest 0.1 mm using an ocular micrometer. Multiple sections of the tumor were studied to identify the area with maximum thickness. The tumor thickness was measured from the level of adjacent normal mucosa to the deepest point of tumor invasion as described by Moore et al. [15] and depicted in Fig. 1. The distance of the tumor from its closest resection margin was measured to the nearest 0.1 mm using an ocular micrometer. The location of the closest margin was not consistently recorded in the database, thus precluding analysis. Other histopathologic factors such as tumor differentiation, patterns of invasion, PNI, and lymphovascular involvement (LVI), lymph node involvement with or without extracapsular spread (ECS) were evaluated as per the College of American Pathologist’s criteria [16]. Statistical analysis Local control (LC) was calculated from the date of surgery to the date of last follow up or local recurrence. Disease specific survival (DSS) was calculated from the date of surgery to the date of last follow up or death from OSCC, with patients dying from other causes being censored at the time of death. Margin to size ratio (MSR) and margin to thickness ratio (MTR) were calculated by dividing the tumor margin in millimeters by tumor diameter (MSR) or tumor thickness (MTR) in millimeters. In cases where the closest margin was greater than 5 mm, a margin of 5.01 mm was imputed because margins more than 5 mm were not routinely recorded. In cases with involved margins, a value of 0.01 mm mm was imputed. The distribution for both MTR and MSR was strongly skewed; hence a natural logarithm transformation was performed. As log (0) cannot be calculated, 1 was added to the margin to allow transformation. Log(MSR) and log(MTR) were analyzed as predictors of
Fig. 1. Two schematized cross sections of OSCC tumors; tumor thickness was measured as the distance from the level of the mucosa (dotted line) to the deepest extent of the tumor. The closest margin in this assumption is between the tumor and the resection (dashed line). In the first example the MTR is 1.5 mm/3 mm = 0.5; in the second example the MTR is 1.5 mm/6 mm = 0.25. Hence, a safe margin for a tumor with a thickness of 4 mm requires 1.2 mm, a safe margin for a tumor with 8 mm thickness would be 2.4 mm and a tumor with a thickness of 15 would require the traditional margin of 5 mm.
LC and survival in isolation and adjusting for the effect of other clinically significant variables such as PNI, LVI, tumor diameter, tumor thickness, lymph node status, ECS, and radiotherapy to ascertain whether log(MSR) and log(MTR) are independent predictors of LC or survival. Each factor was analyzed in three ways: (1) all patients, (2) excluding patients with involved margins, (3) excluding patients with involved and clear (>5 mm) margins. MSR and MRT were then divided into 10 evenly distributed categories (deciles) in order to identify a clinically useful value that could be used in routine practice. In order to determine whether MTR/MSR was a better prognostic factor than margin alone, the Akaike information criterion (AIC) was used which takes into account how well the model fits the data and the complexity of the model. Statistical analysis was performed using SPSS version 22.0 (IBM, Armonk, NY).
Results Cohort characteristics This cohort includes 539 cases of OSCC from 1987–2014. Of these, 38 were excluded due to insufficient data. The remaining 501 patients were included in the statistical analysis. The baseline characteristics of our study population are depicted in Table 1. Median follow-up was 2.3 years (range 0.1–18.6). The association of various clinicopathologic factors with DSS and LC are summarized in Table 2. There were 55 local failures and 108 total deaths, including 60 deaths due to OSCC. Local failures are further broken down in Table 3.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
3
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx Table 1 Summary of baseline patient characteristics.
Table 3 Breakdown of local control statistics by margin and margin to tumor thickness ratio. N (%)/Median [IQR]
Variable
Local recurrence events
2-Year local controla
Age
63.6 [53.2–72.9]
Gender Male Female
289 (57.7) 212 (42.3)
MTR < 0.06 MTR 0.06–0.3 MTR > 0.3 Margin 0.1–2 mm Margin > 2 mm
21/150 18/151 16/200 24/211 19/214
82.9 ± 3.9 86.5 ± 3.3 94.1 ± 1.8 88.6 ± 2.6 91.0 ± 2.2
Maximum tumor diameter (mm)
24.0 [15.0–33.0]
Tumor differentiation Well Moderate Poor
81/465 (17.4) 307/465 (66.0) 77/465 (16.6)
Perineural invasion Lymphovascular invasion
151 (30.1) 69 (13.8)
T stage T1 T2 T3 T4
170 (33.9) 180 (35.9) 43 (8.6) 108 (21.6)
N stage N0 N1 N2 N3
303 (60.5) 56 (11.2) 140 (27.9) 2 (0.4)
Extracapsular spread Radiation Chemotherapy
83 (16.6) 225 (44.9) 29 (5.8)
a Data reported as cumulative disease-free survival ± standard error; MTR, margin to tumor thickness ratio.
as the estimates were similar for all three groups as shown in Table 4. Log(MSR) was a predictor of both LC (HR 0.76; 95% CI, 0.59–0.99; p = 0.040) and DSS (HR 0.73; 95% CI, 0.57–0.94; p = 0.014) on univariable analysis. However, log(MSR) was not a significant predictor of LC (HR 0.74; 95% CI, 0.53–1.05; p = 0.089) on multivariable analysis adjusting for the effects of PNI, LVI, tumor thickness, and radiotherapy. Also, log(MSR) was not a significant predictor of DSS (HR 0.83, 95% CI, 0.63–1.10; p = 0.205) after adjusting for the effects of PNI, LVI, nodal status, ECS, and radiotherapy. On univariable analysis, log(MTR) was a predictor of both LC (HR 0.76; 95% CI, 0.61–0.95; p = 0.017) and DSS (HR 0.70; 95% CI, 0.57–0.87; p = 0.001). Log(MTR) remained a significant predictor of LC (HR 0.71; 95% CI, 0.52–0.97; p = 0.033) after adjusting for the effect of PNI, LVI, tumor diameter, and radiotherapy, in contrast to MSR. Log(MTR) was also a significant predictor of DSS (HR 0.77; 95% CI, 0.60–0.99; p = 0.038) after adjusting for the effect of PNI, LVI, nodal status, ECS, and radiotherapy as shown in Table 5.
Analysis of MSR and MTR as continuous variables Analysis of MTR as a categorical variable Association of log(MSR) and log(MTR) with LC and DSS was initially evaluated using three patient margin groups (all, clear and close, close alone). The entire cohort was used for further analysis,
As MTR was the only significant independent predictor of both LC and DSS, MSR was excluded from further analysis. MTR was
Table 2 Association of clinicopathologic features with disease-specific survival and local control (includes all patients). Factor
Disease-specific survival
Local control
HR (95% CI)
p-value
HR (95% CI)
p-value
Maximum tumor diameter (per mm)
1.02 (1.01–1.04)
0.003
1.01 (0.99–1.03)
0.353
Maximum tumor diameter (categorical) <2 cm (reference) 2–4 cm >4 cm
– 1.85 (0.99–3.43) 3.84 (1.73–8.49)
– 0.053 0.001
– 1.00 (0.56–1.78) 1.91 (0.84–4.35)
– 0.994 0.124
Tumor thickness (per mm)
1.04 (1.02–1.06)
<0.001
1.01 (0.99–1.04)
0.309
Tumor thickness (categorical) 62 mm (reference) >2, <5 mm 5–10 mm >10 mm
– 1.19 (0.23–6.13) 2.81 (0.67–11.86) 3.63 (0.87–15.25)
– 0.837 0.160 0.078
– 1.05 (0.32–3.41) 1.09 (0.37–3.20) 1.39 (0.48–4.04)
– 0.938 0.873 0.543
Tumor margin (per mm)
0.89 (0.77–1.02)
0.094
0.82 (0.71–0.96)
0.012
Tumor margin (categorical) 0 mm (i.e. involved) (reference) 62 mm >2 mm
– 0.78 (0.39–1.59) 0.64 (0.31–1.31)
– 0.498 0.221
– 0.64 (0.32–1.28) 0.45 (0.22–0.92)
– 0.207 0.030
Perineural invasion
2.00 (1.19–3.37)
0.009
2.00 (1.16–3.45)
0.013
Lymphovascular invasion
2.03 (1.06–3.91)
0.034
0.86 (0.34–2.15)
0.738
Pathological T stage T1 (reference) T2 T3 T4
– 1.60 (0.80–3.22) 4.09 (1.74–9.60) 3.05 (1.49–6.23)
– 0.186 0.001 0.002
– 0.76 (0.40–1.47) 1.69 (0.68–4.20) 1.18 (0.58–2.39)
– 0.421 0.259 0.657
Nodal involvement
1.86 (1.12–3.09)
0.016
0.94 (0.54–1.64)
0.826
Extracapsular spread
2.19 (1.20–3.98)
0.011
1.32 (0.65–2.71)
0.443
Radiotherapy
1.83 (1.10–3.06)
0.021
1.05 (0.62–1.79)
0.854
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
4
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
Table 4 Univariable Cox regression for local control and disease-specific survival. Outcome
Cohort
Factor
HR (95% CI)
p-value
Local recurrence
ALL
log(MTR) log(MSR) log(MTR) log(MSR) log(MTR) log(MSR)
0.76 0.76 0.78 0.78 0.89 0.87
(0.61–0.95) (0.59–0.99) (0.59–1.04) (0.57–1.08) (0.65–1.21) (0.61–1.24)
0.017 0.040 0.086 0.135 0.441 0.440
log(MTR) log(MSR) log(MTR) log(MSR) log(MTR) log(MSR)
0.70 0.73 0.69 0.74 0.72 0.77
(0.57–0.87) (0.57–0.94) (0.53–0.90) (0.54–1.00) (0.53–0.97) (0.54–1.10)
0.001 0.014 0.006 0.051 0.032 0.152
Clear and close Close only Disease-specific survival
ALL Clear and close Close only
Table 5 Multivariable Cox regression for disease-specific survival and local control. Factor
HR (95% CI)
p-value
Analyzed as continuous variable Disease specific survival Radiation PNI LVI ECS Nodal stage positive Log(MTR)
1.00 1.42 1.44 1.37 1.20 0.77
(0.53–1.8) (0.82–2.48) (0.72–2.87) (0.65–2.86) (0.63–2.28) (0.60–0.99)
0.990 0.216 0.306 0.408 0.584 0.038
Local control Radiation PNI LVI Maximum diameter (mm) Log(MTR)
0.68 1.84 0.72 0.99 0.71
(0.37–1.26) (1.03–3.32) (0.28–1.84) (0.97–1.02) (0.52–0.97)
0.216 0.041 0.488 0.612 0.033
Analyzed as categorial variable PNI LVI Nodal stage positive
1.55 (0.89–2.71) 1.56 (0.78–3.10) 1.36 (0.79–2.35)
0.124 0.207 0.273
MTR (>0.3 as reference) <0.06 0.06–0.3
1.92 (1.02–3.61) 1.22 (0.62–2.41)
0.043 0.559
divided into 10 equal groups (deciles) as shown in Fig. 2A and B. A progressive increase in hazard was observed with decrease in the MTR. MTR was divided into three groups (decile 1–3 (<0.06), 4–6 (0.06–0.3) and 7–10 (>0.3) as shown in Fig. 2C and D). MTR was then analyzed using these three categories in the multivariable model as shown in Table 5. As MTR is most applicable in patients with close margins rather than those with involved margins, where the MTR will always equal zero, we checked the model after exclusion of 79 patients with an involved margin. The MTR < 0.06 remained significant in the multivariable model for both LC (HR 2.64, CI 1.05–6.63, p = 0.039) and DSS (HR 2.27, CI 1.07–4.81, p = 0.032) with over twice the risk of local recurrence and disease specific death as compared to a MTR > 0.3. Comparison of MTR with margin groups To compare the MTR with classical margins, we divided our cohort into patients with 62 mm and >2 mm margins. LC and DSS were calculated for these subgroups (Table 3). The Akaike information criterion (AIC) was calculated for univariable and multivariable models including MTR and margin 62 mm v >2 mm. Smaller AICs were observed for all models using MTR compared to models using 2 mm margins indicating that MTR is a superior predictor of LC and DSS, (data not shown). There were 106 patients with a margin of >5 mm, of which 7 patients had a local recurrence
with a 2-year LC of 94.6% ± 2.4%. In comparison there were 200 patients with an MTR > 0.3 with a comparable number of local recurrences (16 patients) and LC rate (94.1% ± 1.8%).
Discussion We report on a retrospectively analyzed group of OSCC patients treated in our institution over a period of 27 years and demonstrate that MTR is a significant predictor of both LC and DSS. In fact, MTR remained a significant predictor of recurrence and survival in our models after adjusting for important prognostic factors such as nodal status and ECS. MTR performed better in all models and was more reproducible than MSR, and hence appears to be a more reliable predictor. The concept of MTR is useful as it combines a measure of the volume with the estimated margin. It is important to consider that the ‘true’ closest margin is unknown because the pathologist cannot evaluate every cell of the tumor/normal tissue interface. Hence one would expect the uncertainty to increase with increasing tumor volume. The relationship between tumor thickness and resection margins is well established in melanoma where the suggested MTR is much higher than what we have found for OSCC, for example for a 1 mm thick melanoma a resection margin of 10 mm is recommended (MTR = 10). A ratio has also been reported in non small cell lung cancer, with an MSR larger than 1 suggested to be safe [17]. MTR > 0.3 has considerable clinical utility, particularly in thin tumors (thickness <4 mm) (Fig. 1) where in most cases a surgical margin of 1–2 mm would still be considered as low risk based on this data. The clinical utility of MTR in very thick tumors (>10 mm) is probably limited to rare examples of well differentiated cancers with pushing margins and no other adverse features, which may potentially be managed with surgery alone when the MTR is >0.3. Our data indicate that a value of <0.06 is associated with more than twice the risk of recurrence and death after adjusting for the effects of other clinically important variables such as PNI, LVI, and radiotherapy compared to an MTR > 0.3. An MTR < 0.06 represents a very high risk value, and could be used to define what is effectively a positive margin from the perspective of being an indication for adjuvant chemoradiation. From a clinical point of view, the safe margin for a tumor without other adverse features can be calculated by multiplying the tumor thickness by 0.3 to obtain the respective value. Examples for various tumor thicknesses are given in Fig. 1. In our cohort, 106 patients had a margin of >5 mm, which would traditionally be considered sufficient. In contrast, the number of patients with a ‘‘safe” MTR > 0.3 was 200. The LC and DSS was similar in both groups; therefore by using MTR potentially more patients can be spared adjuvant therapy. MTR performed
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
5
Fig. 2. Survival curves according to margin/tumor thickness ratio (MTR). Local recurrence free survival (A) and disease specific survival (B) for MTR was divided into 10 deciles, showing a decreased survival with reducing MTR. Subgroups were consolidated for local recurrence free survival (C) and disease specific survival (D). Group 1 included deciles 1–3 (MTR < 0.06), group 2 included deciles 4–6 (0.06 < MTR < 0.3) and group 3 included deciles 7–10 (MTR > 0.3).
better than any traditional margin (2 mm or 5 mm) as a predictor of disease control in the models constructed. It is reassuring that MTR has been a consistent predictor of both LC and DSS as both a continuous and categorical variable in all cohorts of patients, i.e. with close margins and all patients combined. While it is unrealistic to expect particular cut-offs from one institution to be applicable to others, we expect the principle to be broadly replicable. We assume that MTR needs to be considered as only one of many possible risk factors, such as LNR [9], PNI [18], or LVI [19]. The important institutional confounding variables that cannot be adjusted for are specific treatment effects (type of surgery, dose and distribution of radiotherapy, chemotherapy agents and dose) and pathological technique in assessing margins. For the purposes of making MTR easy to test, we suspect that a MTR < 0.1 (very high risk) and an MTR > 0.3 (low risk) would be reasonable values for other institutions to validate, using their own data. Conflict of interest statement No conflict of interest. Acknowledgement No financial support was provided for this study.
References [1] Helliwell T, Julia W. Standards and datasets for reporting cancers. Dataset for histopathology reporting of mucosal malignancies of the oral cavity; 2013. [2] Alicandri-Ciufelli M, Bonali M, Piccinini A, Marra L, Ghidini A, Cunsolo EM, et al. Surgical margins in head and neck squamous cell carcinoma: what is ‘‘close?”. Eur Arch Otorhinolaryngol 2012;270:2603–9. http://dx.doi.org/10.1007/ s00405-012-2317-8. [3] Dixit S, Vyas RK, Toparani RB, Baboo HA, Patel DD. Surgery versus surgery and postoperative radiotherapy in squamous cell carcinoma of the buccal mucosa: a comparative study. Ann Surg Oncol 1998;5:502–10. [4] Liao C-T, Huang S-F, Chen I-H, Chang JT-C, Wang H-M, Ng S-H, et al. When does skin excision allow the achievement of an adequate local control rate in patients with squamous cell carcinoma involving the buccal mucosa? Ann Surg Oncol 2008;15:2187–94. http://dx.doi.org/10.1245/s10434-008-9980-4. [5] Ch’ng S, Corbett-Burns S, Stanton N, Gao K, Shannon K, Clifford A, et al. Close margin alone does not warrant postoperative adjuvant radiotherapy in oral squamous cell carcinoma. Cancer 2013;119:2427–37. http://dx.doi.org/ 10.1002/cncr.28081. [6] Brandwein-Gensler M, Teixeira MS, Lewis CM, Lee B, Rolnitzky L, Hille JJ, et al. Oral squamous cell carcinoma: histologic risk assessment, but not margin status, is strongly predictive of local disease-free and overall survival. Am J Surg Pathol 2005;29:167–78. [7] International Consortium for Outcome Research (ICOR) in Head and Neck Cancer , Ebrahimi A, Gil Z, Amit M, Yen T-C, Liao C-T, et al. Primary tumor staging for oral cancer and a proposed modification incorporating depth of invasion: an international multicenter retrospective study. JAMA Otolaryngol Head Neck Surg 2014;140:1138–48. http://dx.doi.org/ 10.1001/jamaoto.2014.1548. [8] Ong W, Zhao R, Lui B, Tan W, Ebrahimi A, Clark JR, et al. Prognostic significance of lymph node density in squamous cell carcinoma of the tongue. Head Neck 2015. http://dx.doi.org/10.1002/hed.24113.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
6
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
[9] Ebrahimi A, Clark JR, Zhang WJ, Elliott MS, Gao K, Milross CG, et al. Lymph node ratio as an independent prognostic factor in oral squamous cell carcinoma. Head Neck 2011;33:1245–51. http://dx.doi.org/10.1002/hed.21600. [10] Fong ZV, Tanabe KK. Comparison of melanoma guidelines in the U.S.A., Canada, Europe, Australia and New Zealand: a critical appraisal and comprehensive review. Br J Dermatol 2014;170:20–30. http://dx.doi.org/10.1111/bjd.12687. [11] Nathanson A, Agren K, Biörklund A, Lind MG, Andréason L, Anniko M, et al. Evaluation of some prognostic factors in small squamous cell carcinoma of the mobile tongue: a multicenter study in Sweden. Head Neck 1989;11:387–92. [12] Yuen AP, Lam KY, Wei WI, Ho CM, Chow TL, Yuen WF. A comparison of the prognostic significance of tumor diameter, length, width, thickness, area, volume, and clinicopathological features of oral tongue carcinoma. Am J Surg 2000;180:139–43. [13] Ichimiya Y, Fuwa N, Kamata M, Kodaira T, Furutani K, Tachibana H, et al. Treatment results of stage I oral tongue cancer with definitive radiotherapy. Oral Oncol 2005;41:520–5. http://dx.doi.org/10.1016/j.oraloncology.2004. 12.012. [14] Balasubramanian D, Ebrahimi A, Gupta R, Gao K, Elliott M, Palme CE, et al. Tumour thickness as a predictor of nodal metastases in oral cancer:
[15] [16]
[17]
[18]
[19]
comparison between tongue and floor of mouth subsites. Oral Oncol 2014;50:1165–8. http://dx.doi.org/10.1016/j.oraloncology.2014.09.012. Moore C, Kuhns JG, Greenberg RA. Thickness as prognostic aid in upper aerodigestive tract cancer. Arch Surg 1986;121:1410–4. Seethala R, Ilan W, Diane C, McHugh JB, Harrison LB, Richardson MS, et al. Protocol for the examination of specimens from patients with carcinomas of the lip and oral cavity; 2015. Sawabata N, Maeda H, Matsumura A, Ohta M, Okumura M. Thoracic Surgery Study Group of Osaka University. Clinical implications of the margin cytology findings and margin/tumor size ratio in patients who underwent pulmonary excision for peripheral non-small cell lung cancer. Surg Today 2012;42:238–44. http://dx.doi.org/10.1007/s00595-011-0031-6. Aivazian K, Ebrahimi A, Low T-HH, Gao K, Clifford A, Shannon K, et al. Perineural invasion in oral squamous cell carcinoma: quantitative subcategorisation of perineural invasion and prognostication. J Surg Oncol 2015;111:352–8. http://dx.doi.org/10.1002/jso.23821. Tai S-K, Li W-Y, Chu P-Y, Chang S-Y, Tsai T-L, Wang Y-F, et al. Risks and clinical implications of perineural invasion in T1-2 oral tongue squamous cell carcinoma. Head Neck 2012;34:994–1001. http://dx.doi.org/10.1002/ hed.21846.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
Contents lists available at ScienceDirect
Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology
Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma Gregor Heiduschka a,b,⇑, Sohaib A. Virk c, Carsten E. Palme a, Sydney Ch’ng a,g, Michael Elliot a, Ruta Gupta d, Jonathan Clark a,e,f a
Department of Head and Neck Surgery, Chris O’Brien Lifehouse, Camperdown, New South Wales, Australia Department of Otorhinolaryngology – Head and Neck Surgery, Medical University of Vienna, Austria University of New South Wales, Randwick, New South Wales, Australia d Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia e Central Clinical School, The University of Sydney, Sydney, New South Wales, Australia f South Western Sydney Clinical School, University of New South Wales, Sydney, New South Wales, Australia g The Institute of Academic Surgery at RPA, University of Sydney, NSW, Australia b c
a r t i c l e
i n f o
Article history: Received 8 September 2015 Received in revised form 9 January 2016 Accepted 13 January 2016 Available online xxxx Keywords: OSCC Margin Thickness Prognostic marker
s u m m a r y Objectives: To assess whether small oral squamous cell carcinomas (OSCC) require the same margin clearance as large tumors. We evaluated the association between the ratio of the closest margin to tumor size (MSR) and tumor thickness (MTR) with local control and survival. Methods and methods: The clinicopathologic and follow up data were obtained for 501 OSCC patients who had surgical resection with curative intent at our institution. MTR and MSR were computed and their associations with local control and survival were assessed using multivariable Cox-regression model. Survival curves were generated using the Kaplan–Meier method. Results: MTR was a better predictor of disease control than MSR. MTR was a predictor of local failure (p = 0.033) and disease specific death (p = 0.038) after adjusting for perineural invasion, lymphovascular involvement, nodal status, and radiotherapy. A threshold MTR value of 0.3 was identified, above which the risk of local recurrence was low. Conclusion: The ratio of margin to tumor thickness was an independent predictor for local recurrence and disease specific death in this cohort. A MTR > 0.3 can serve as a useful tool for adjuvant therapy planning as it combines tumor thickness and margin clearance, two well established prognostic factors. The minimum safe margin can be calculated by multiplying the tumor thickness by 0.3. Further prospective studies in other institutions are warranted to confirm the prognostic utility of MTR and assess the generalizability of our threshold values. Ó 2016 Elsevier Ltd. All rights reserved.
Introduction Surgical resection is the mainstay of treatment for oral squamous cell carcinoma (OSCC). Satisfactory oncologic resections require negative and adequate resection margins. A microscopic margin of 5 mm has traditionally been accepted as adequate for various oral cavity sub-sites [1,2]. This definition, however, of adequate surgical resection remains controversial as some authors propose 2 mm to be sufficient [3] while others advocate for more than 7 mm [4]. Ch’ng et al. [5] and Brandwein-Gensler et al. [6] ⇑ Corresponding author at: Sydney Head & Neck Cancer Institute, The Chris O’Brien Lifehouse, Royal Prince Alfred Hospital, Missenden Road, Camperdown, NSW 2050, Australia. Tel.: +61 8514 0109; fax: +61 9519 9214. E-mail address: [email protected] (G. Heiduschka).
have suggested that the adequacy of surgical resections need to be considered in the context of various pathological characteristics such as differentiation, growth pattern, and perineural invasion (PNI). It is intuitive that larger tumors with an infiltrative growth pattern would require larger margins compared to smaller tumors, particularly those with pushing margins. Thus a ratio that takes into account well established prognostic factors such as tumor size, tumor thickness, and distance to the closest margin may help evaluate whether a certain margin can be considered as safe. The clinical utility of ratios of two related pathological variables is becoming increasingly recognized in oncology. An important example of this is lymph node ratio (LNR) which is a strong predictor of survival in oral cancer [7–9]. LNR incorporates not only the number of lymph node metastases, but also the total number of nodes counted by the pathologist, which enables information on
http://dx.doi.org/10.1016/j.oraloncology.2016.01.010 1368-8375/Ó 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
2
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
burden of disease, comprehensiveness of surgical resection, and thoroughness of pathological examination to be combined into a single parameter. Similar guidelines regarding adequate surgical resections are well established in melanoma based on depth of invasion [10]. Although tumor thickness and depth of invasion have been shown to be important prognostic factors in OSCC, the relation between tumor size or thickness and the resection margins has been not been studied [11–14]. The aim of this study is to evaluate whether a ratio between the resection margin and tumor thickness or tumor size is a predictor of local control or survival in OSCC and whether a clinically useful threshold value that may aid adjuvant therapy planning can be identified. Methods Study population The Sydney Head and Neck Cancer Institute has maintained a prospective database including the clinical, pathologic and follow-up data of all patients treated in the Department of Head and Neck Surgery, Royal Prince Alfred Hospital and Chris O’Brien Lifehouse since October 1987. After obtaining institutional ethics approval, clinico-pathological data for all patients with OSCC treated between October 1987 and December 2014 was extracted from the database. Histologic variables Tumor size was defined as the largest dimension of the tumor as measured in millimeters during macroscopic examination and confirmed microscopically for tumors less than 10 mm in maximum dimension. Tumor thickness was measured on formalin fixed paraffin embedded sections stained with haematoxylin and eosin to the nearest 0.1 mm using an ocular micrometer. Multiple sections of the tumor were studied to identify the area with maximum thickness. The tumor thickness was measured from the level of adjacent normal mucosa to the deepest point of tumor invasion as described by Moore et al. [15] and depicted in Fig. 1. The distance of the tumor from its closest resection margin was measured to the nearest 0.1 mm using an ocular micrometer. The location of the closest margin was not consistently recorded in the database, thus precluding analysis. Other histopathologic factors such as tumor differentiation, patterns of invasion, PNI, and lymphovascular involvement (LVI), lymph node involvement with or without extracapsular spread (ECS) were evaluated as per the College of American Pathologist’s criteria [16]. Statistical analysis Local control (LC) was calculated from the date of surgery to the date of last follow up or local recurrence. Disease specific survival (DSS) was calculated from the date of surgery to the date of last follow up or death from OSCC, with patients dying from other causes being censored at the time of death. Margin to size ratio (MSR) and margin to thickness ratio (MTR) were calculated by dividing the tumor margin in millimeters by tumor diameter (MSR) or tumor thickness (MTR) in millimeters. In cases where the closest margin was greater than 5 mm, a margin of 5.01 mm was imputed because margins more than 5 mm were not routinely recorded. In cases with involved margins, a value of 0.01 mm mm was imputed. The distribution for both MTR and MSR was strongly skewed; hence a natural logarithm transformation was performed. As log (0) cannot be calculated, 1 was added to the margin to allow transformation. Log(MSR) and log(MTR) were analyzed as predictors of
Fig. 1. Two schematized cross sections of OSCC tumors; tumor thickness was measured as the distance from the level of the mucosa (dotted line) to the deepest extent of the tumor. The closest margin in this assumption is between the tumor and the resection (dashed line). In the first example the MTR is 1.5 mm/3 mm = 0.5; in the second example the MTR is 1.5 mm/6 mm = 0.25. Hence, a safe margin for a tumor with a thickness of 4 mm requires 1.2 mm, a safe margin for a tumor with 8 mm thickness would be 2.4 mm and a tumor with a thickness of 15 would require the traditional margin of 5 mm.
LC and survival in isolation and adjusting for the effect of other clinically significant variables such as PNI, LVI, tumor diameter, tumor thickness, lymph node status, ECS, and radiotherapy to ascertain whether log(MSR) and log(MTR) are independent predictors of LC or survival. Each factor was analyzed in three ways: (1) all patients, (2) excluding patients with involved margins, (3) excluding patients with involved and clear (>5 mm) margins. MSR and MRT were then divided into 10 evenly distributed categories (deciles) in order to identify a clinically useful value that could be used in routine practice. In order to determine whether MTR/MSR was a better prognostic factor than margin alone, the Akaike information criterion (AIC) was used which takes into account how well the model fits the data and the complexity of the model. Statistical analysis was performed using SPSS version 22.0 (IBM, Armonk, NY).
Results Cohort characteristics This cohort includes 539 cases of OSCC from 1987–2014. Of these, 38 were excluded due to insufficient data. The remaining 501 patients were included in the statistical analysis. The baseline characteristics of our study population are depicted in Table 1. Median follow-up was 2.3 years (range 0.1–18.6). The association of various clinicopathologic factors with DSS and LC are summarized in Table 2. There were 55 local failures and 108 total deaths, including 60 deaths due to OSCC. Local failures are further broken down in Table 3.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
3
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx Table 1 Summary of baseline patient characteristics.
Table 3 Breakdown of local control statistics by margin and margin to tumor thickness ratio. N (%)/Median [IQR]
Variable
Local recurrence events
2-Year local controla
Age
63.6 [53.2–72.9]
Gender Male Female
289 (57.7) 212 (42.3)
MTR < 0.06 MTR 0.06–0.3 MTR > 0.3 Margin 0.1–2 mm Margin > 2 mm
21/150 18/151 16/200 24/211 19/214
82.9 ± 3.9 86.5 ± 3.3 94.1 ± 1.8 88.6 ± 2.6 91.0 ± 2.2
Maximum tumor diameter (mm)
24.0 [15.0–33.0]
Tumor differentiation Well Moderate Poor
81/465 (17.4) 307/465 (66.0) 77/465 (16.6)
Perineural invasion Lymphovascular invasion
151 (30.1) 69 (13.8)
T stage T1 T2 T3 T4
170 (33.9) 180 (35.9) 43 (8.6) 108 (21.6)
N stage N0 N1 N2 N3
303 (60.5) 56 (11.2) 140 (27.9) 2 (0.4)
Extracapsular spread Radiation Chemotherapy
83 (16.6) 225 (44.9) 29 (5.8)
a Data reported as cumulative disease-free survival ± standard error; MTR, margin to tumor thickness ratio.
as the estimates were similar for all three groups as shown in Table 4. Log(MSR) was a predictor of both LC (HR 0.76; 95% CI, 0.59–0.99; p = 0.040) and DSS (HR 0.73; 95% CI, 0.57–0.94; p = 0.014) on univariable analysis. However, log(MSR) was not a significant predictor of LC (HR 0.74; 95% CI, 0.53–1.05; p = 0.089) on multivariable analysis adjusting for the effects of PNI, LVI, tumor thickness, and radiotherapy. Also, log(MSR) was not a significant predictor of DSS (HR 0.83, 95% CI, 0.63–1.10; p = 0.205) after adjusting for the effects of PNI, LVI, nodal status, ECS, and radiotherapy. On univariable analysis, log(MTR) was a predictor of both LC (HR 0.76; 95% CI, 0.61–0.95; p = 0.017) and DSS (HR 0.70; 95% CI, 0.57–0.87; p = 0.001). Log(MTR) remained a significant predictor of LC (HR 0.71; 95% CI, 0.52–0.97; p = 0.033) after adjusting for the effect of PNI, LVI, tumor diameter, and radiotherapy, in contrast to MSR. Log(MTR) was also a significant predictor of DSS (HR 0.77; 95% CI, 0.60–0.99; p = 0.038) after adjusting for the effect of PNI, LVI, nodal status, ECS, and radiotherapy as shown in Table 5.
Analysis of MSR and MTR as continuous variables Analysis of MTR as a categorical variable Association of log(MSR) and log(MTR) with LC and DSS was initially evaluated using three patient margin groups (all, clear and close, close alone). The entire cohort was used for further analysis,
As MTR was the only significant independent predictor of both LC and DSS, MSR was excluded from further analysis. MTR was
Table 2 Association of clinicopathologic features with disease-specific survival and local control (includes all patients). Factor
Disease-specific survival
Local control
HR (95% CI)
p-value
HR (95% CI)
p-value
Maximum tumor diameter (per mm)
1.02 (1.01–1.04)
0.003
1.01 (0.99–1.03)
0.353
Maximum tumor diameter (categorical) <2 cm (reference) 2–4 cm >4 cm
– 1.85 (0.99–3.43) 3.84 (1.73–8.49)
– 0.053 0.001
– 1.00 (0.56–1.78) 1.91 (0.84–4.35)
– 0.994 0.124
Tumor thickness (per mm)
1.04 (1.02–1.06)
<0.001
1.01 (0.99–1.04)
0.309
Tumor thickness (categorical) 62 mm (reference) >2, <5 mm 5–10 mm >10 mm
– 1.19 (0.23–6.13) 2.81 (0.67–11.86) 3.63 (0.87–15.25)
– 0.837 0.160 0.078
– 1.05 (0.32–3.41) 1.09 (0.37–3.20) 1.39 (0.48–4.04)
– 0.938 0.873 0.543
Tumor margin (per mm)
0.89 (0.77–1.02)
0.094
0.82 (0.71–0.96)
0.012
Tumor margin (categorical) 0 mm (i.e. involved) (reference) 62 mm >2 mm
– 0.78 (0.39–1.59) 0.64 (0.31–1.31)
– 0.498 0.221
– 0.64 (0.32–1.28) 0.45 (0.22–0.92)
– 0.207 0.030
Perineural invasion
2.00 (1.19–3.37)
0.009
2.00 (1.16–3.45)
0.013
Lymphovascular invasion
2.03 (1.06–3.91)
0.034
0.86 (0.34–2.15)
0.738
Pathological T stage T1 (reference) T2 T3 T4
– 1.60 (0.80–3.22) 4.09 (1.74–9.60) 3.05 (1.49–6.23)
– 0.186 0.001 0.002
– 0.76 (0.40–1.47) 1.69 (0.68–4.20) 1.18 (0.58–2.39)
– 0.421 0.259 0.657
Nodal involvement
1.86 (1.12–3.09)
0.016
0.94 (0.54–1.64)
0.826
Extracapsular spread
2.19 (1.20–3.98)
0.011
1.32 (0.65–2.71)
0.443
Radiotherapy
1.83 (1.10–3.06)
0.021
1.05 (0.62–1.79)
0.854
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
4
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
Table 4 Univariable Cox regression for local control and disease-specific survival. Outcome
Cohort
Factor
HR (95% CI)
p-value
Local recurrence
ALL
log(MTR) log(MSR) log(MTR) log(MSR) log(MTR) log(MSR)
0.76 0.76 0.78 0.78 0.89 0.87
(0.61–0.95) (0.59–0.99) (0.59–1.04) (0.57–1.08) (0.65–1.21) (0.61–1.24)
0.017 0.040 0.086 0.135 0.441 0.440
log(MTR) log(MSR) log(MTR) log(MSR) log(MTR) log(MSR)
0.70 0.73 0.69 0.74 0.72 0.77
(0.57–0.87) (0.57–0.94) (0.53–0.90) (0.54–1.00) (0.53–0.97) (0.54–1.10)
0.001 0.014 0.006 0.051 0.032 0.152
Clear and close Close only Disease-specific survival
ALL Clear and close Close only
Table 5 Multivariable Cox regression for disease-specific survival and local control. Factor
HR (95% CI)
p-value
Analyzed as continuous variable Disease specific survival Radiation PNI LVI ECS Nodal stage positive Log(MTR)
1.00 1.42 1.44 1.37 1.20 0.77
(0.53–1.8) (0.82–2.48) (0.72–2.87) (0.65–2.86) (0.63–2.28) (0.60–0.99)
0.990 0.216 0.306 0.408 0.584 0.038
Local control Radiation PNI LVI Maximum diameter (mm) Log(MTR)
0.68 1.84 0.72 0.99 0.71
(0.37–1.26) (1.03–3.32) (0.28–1.84) (0.97–1.02) (0.52–0.97)
0.216 0.041 0.488 0.612 0.033
Analyzed as categorial variable PNI LVI Nodal stage positive
1.55 (0.89–2.71) 1.56 (0.78–3.10) 1.36 (0.79–2.35)
0.124 0.207 0.273
MTR (>0.3 as reference) <0.06 0.06–0.3
1.92 (1.02–3.61) 1.22 (0.62–2.41)
0.043 0.559
divided into 10 equal groups (deciles) as shown in Fig. 2A and B. A progressive increase in hazard was observed with decrease in the MTR. MTR was divided into three groups (decile 1–3 (<0.06), 4–6 (0.06–0.3) and 7–10 (>0.3) as shown in Fig. 2C and D). MTR was then analyzed using these three categories in the multivariable model as shown in Table 5. As MTR is most applicable in patients with close margins rather than those with involved margins, where the MTR will always equal zero, we checked the model after exclusion of 79 patients with an involved margin. The MTR < 0.06 remained significant in the multivariable model for both LC (HR 2.64, CI 1.05–6.63, p = 0.039) and DSS (HR 2.27, CI 1.07–4.81, p = 0.032) with over twice the risk of local recurrence and disease specific death as compared to a MTR > 0.3. Comparison of MTR with margin groups To compare the MTR with classical margins, we divided our cohort into patients with 62 mm and >2 mm margins. LC and DSS were calculated for these subgroups (Table 3). The Akaike information criterion (AIC) was calculated for univariable and multivariable models including MTR and margin 62 mm v >2 mm. Smaller AICs were observed for all models using MTR compared to models using 2 mm margins indicating that MTR is a superior predictor of LC and DSS, (data not shown). There were 106 patients with a margin of >5 mm, of which 7 patients had a local recurrence
with a 2-year LC of 94.6% ± 2.4%. In comparison there were 200 patients with an MTR > 0.3 with a comparable number of local recurrences (16 patients) and LC rate (94.1% ± 1.8%).
Discussion We report on a retrospectively analyzed group of OSCC patients treated in our institution over a period of 27 years and demonstrate that MTR is a significant predictor of both LC and DSS. In fact, MTR remained a significant predictor of recurrence and survival in our models after adjusting for important prognostic factors such as nodal status and ECS. MTR performed better in all models and was more reproducible than MSR, and hence appears to be a more reliable predictor. The concept of MTR is useful as it combines a measure of the volume with the estimated margin. It is important to consider that the ‘true’ closest margin is unknown because the pathologist cannot evaluate every cell of the tumor/normal tissue interface. Hence one would expect the uncertainty to increase with increasing tumor volume. The relationship between tumor thickness and resection margins is well established in melanoma where the suggested MTR is much higher than what we have found for OSCC, for example for a 1 mm thick melanoma a resection margin of 10 mm is recommended (MTR = 10). A ratio has also been reported in non small cell lung cancer, with an MSR larger than 1 suggested to be safe [17]. MTR > 0.3 has considerable clinical utility, particularly in thin tumors (thickness <4 mm) (Fig. 1) where in most cases a surgical margin of 1–2 mm would still be considered as low risk based on this data. The clinical utility of MTR in very thick tumors (>10 mm) is probably limited to rare examples of well differentiated cancers with pushing margins and no other adverse features, which may potentially be managed with surgery alone when the MTR is >0.3. Our data indicate that a value of <0.06 is associated with more than twice the risk of recurrence and death after adjusting for the effects of other clinically important variables such as PNI, LVI, and radiotherapy compared to an MTR > 0.3. An MTR < 0.06 represents a very high risk value, and could be used to define what is effectively a positive margin from the perspective of being an indication for adjuvant chemoradiation. From a clinical point of view, the safe margin for a tumor without other adverse features can be calculated by multiplying the tumor thickness by 0.3 to obtain the respective value. Examples for various tumor thicknesses are given in Fig. 1. In our cohort, 106 patients had a margin of >5 mm, which would traditionally be considered sufficient. In contrast, the number of patients with a ‘‘safe” MTR > 0.3 was 200. The LC and DSS was similar in both groups; therefore by using MTR potentially more patients can be spared adjuvant therapy. MTR performed
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
5
Fig. 2. Survival curves according to margin/tumor thickness ratio (MTR). Local recurrence free survival (A) and disease specific survival (B) for MTR was divided into 10 deciles, showing a decreased survival with reducing MTR. Subgroups were consolidated for local recurrence free survival (C) and disease specific survival (D). Group 1 included deciles 1–3 (MTR < 0.06), group 2 included deciles 4–6 (0.06 < MTR < 0.3) and group 3 included deciles 7–10 (MTR > 0.3).
better than any traditional margin (2 mm or 5 mm) as a predictor of disease control in the models constructed. It is reassuring that MTR has been a consistent predictor of both LC and DSS as both a continuous and categorical variable in all cohorts of patients, i.e. with close margins and all patients combined. While it is unrealistic to expect particular cut-offs from one institution to be applicable to others, we expect the principle to be broadly replicable. We assume that MTR needs to be considered as only one of many possible risk factors, such as LNR [9], PNI [18], or LVI [19]. The important institutional confounding variables that cannot be adjusted for are specific treatment effects (type of surgery, dose and distribution of radiotherapy, chemotherapy agents and dose) and pathological technique in assessing margins. For the purposes of making MTR easy to test, we suspect that a MTR < 0.1 (very high risk) and an MTR > 0.3 (low risk) would be reasonable values for other institutions to validate, using their own data. Conflict of interest statement No conflict of interest. Acknowledgement No financial support was provided for this study.
References [1] Helliwell T, Julia W. Standards and datasets for reporting cancers. Dataset for histopathology reporting of mucosal malignancies of the oral cavity; 2013. [2] Alicandri-Ciufelli M, Bonali M, Piccinini A, Marra L, Ghidini A, Cunsolo EM, et al. Surgical margins in head and neck squamous cell carcinoma: what is ‘‘close?”. Eur Arch Otorhinolaryngol 2012;270:2603–9. http://dx.doi.org/10.1007/ s00405-012-2317-8. [3] Dixit S, Vyas RK, Toparani RB, Baboo HA, Patel DD. Surgery versus surgery and postoperative radiotherapy in squamous cell carcinoma of the buccal mucosa: a comparative study. Ann Surg Oncol 1998;5:502–10. [4] Liao C-T, Huang S-F, Chen I-H, Chang JT-C, Wang H-M, Ng S-H, et al. When does skin excision allow the achievement of an adequate local control rate in patients with squamous cell carcinoma involving the buccal mucosa? Ann Surg Oncol 2008;15:2187–94. http://dx.doi.org/10.1245/s10434-008-9980-4. [5] Ch’ng S, Corbett-Burns S, Stanton N, Gao K, Shannon K, Clifford A, et al. Close margin alone does not warrant postoperative adjuvant radiotherapy in oral squamous cell carcinoma. Cancer 2013;119:2427–37. http://dx.doi.org/ 10.1002/cncr.28081. [6] Brandwein-Gensler M, Teixeira MS, Lewis CM, Lee B, Rolnitzky L, Hille JJ, et al. Oral squamous cell carcinoma: histologic risk assessment, but not margin status, is strongly predictive of local disease-free and overall survival. Am J Surg Pathol 2005;29:167–78. [7] International Consortium for Outcome Research (ICOR) in Head and Neck Cancer , Ebrahimi A, Gil Z, Amit M, Yen T-C, Liao C-T, et al. Primary tumor staging for oral cancer and a proposed modification incorporating depth of invasion: an international multicenter retrospective study. JAMA Otolaryngol Head Neck Surg 2014;140:1138–48. http://dx.doi.org/ 10.1001/jamaoto.2014.1548. [8] Ong W, Zhao R, Lui B, Tan W, Ebrahimi A, Clark JR, et al. Prognostic significance of lymph node density in squamous cell carcinoma of the tongue. Head Neck 2015. http://dx.doi.org/10.1002/hed.24113.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010
6
G. Heiduschka et al. / Oral Oncology xxx (2016) xxx–xxx
[9] Ebrahimi A, Clark JR, Zhang WJ, Elliott MS, Gao K, Milross CG, et al. Lymph node ratio as an independent prognostic factor in oral squamous cell carcinoma. Head Neck 2011;33:1245–51. http://dx.doi.org/10.1002/hed.21600. [10] Fong ZV, Tanabe KK. Comparison of melanoma guidelines in the U.S.A., Canada, Europe, Australia and New Zealand: a critical appraisal and comprehensive review. Br J Dermatol 2014;170:20–30. http://dx.doi.org/10.1111/bjd.12687. [11] Nathanson A, Agren K, Biörklund A, Lind MG, Andréason L, Anniko M, et al. Evaluation of some prognostic factors in small squamous cell carcinoma of the mobile tongue: a multicenter study in Sweden. Head Neck 1989;11:387–92. [12] Yuen AP, Lam KY, Wei WI, Ho CM, Chow TL, Yuen WF. A comparison of the prognostic significance of tumor diameter, length, width, thickness, area, volume, and clinicopathological features of oral tongue carcinoma. Am J Surg 2000;180:139–43. [13] Ichimiya Y, Fuwa N, Kamata M, Kodaira T, Furutani K, Tachibana H, et al. Treatment results of stage I oral tongue cancer with definitive radiotherapy. Oral Oncol 2005;41:520–5. http://dx.doi.org/10.1016/j.oraloncology.2004. 12.012. [14] Balasubramanian D, Ebrahimi A, Gupta R, Gao K, Elliott M, Palme CE, et al. Tumour thickness as a predictor of nodal metastases in oral cancer:
[15] [16]
[17]
[18]
[19]
comparison between tongue and floor of mouth subsites. Oral Oncol 2014;50:1165–8. http://dx.doi.org/10.1016/j.oraloncology.2014.09.012. Moore C, Kuhns JG, Greenberg RA. Thickness as prognostic aid in upper aerodigestive tract cancer. Arch Surg 1986;121:1410–4. Seethala R, Ilan W, Diane C, McHugh JB, Harrison LB, Richardson MS, et al. Protocol for the examination of specimens from patients with carcinomas of the lip and oral cavity; 2015. Sawabata N, Maeda H, Matsumura A, Ohta M, Okumura M. Thoracic Surgery Study Group of Osaka University. Clinical implications of the margin cytology findings and margin/tumor size ratio in patients who underwent pulmonary excision for peripheral non-small cell lung cancer. Surg Today 2012;42:238–44. http://dx.doi.org/10.1007/s00595-011-0031-6. Aivazian K, Ebrahimi A, Low T-HH, Gao K, Clifford A, Shannon K, et al. Perineural invasion in oral squamous cell carcinoma: quantitative subcategorisation of perineural invasion and prognostication. J Surg Oncol 2015;111:352–8. http://dx.doi.org/10.1002/jso.23821. Tai S-K, Li W-Y, Chu P-Y, Chang S-Y, Tsai T-L, Wang Y-F, et al. Risks and clinical implications of perineural invasion in T1-2 oral tongue squamous cell carcinoma. Head Neck 2012;34:994–1001. http://dx.doi.org/10.1002/ hed.21846.
Please cite this article in press as: Heiduschka G et al. Margin to tumor thickness ratio – A predictor of local recurrence and survival in oral squamous cell carcinoma. Oral Oncol (2016), http://dx.doi.org/10.1016/j.oraloncology.2016.01.010