There is a broad consensus on the fact that positive margins increase the risk of local recurrence (4–13) and that radiotherapy can improve local control(25–28). When it comes to an adequate resection margin, however, there is still some dissent. This is mostly due to the rarity of soft tissue sarcomas which only account for approximately 1% of new malignant neoplasms in adults. Hence, to reach enough cases, studies about STS often encompass heterogenous patient populations with varied tumor entities, grades and treatments across different institutions and therefore different surgeons, procedures and pathologists. Moreover, they integrate re-resections, whoops procedures, LR, and DM at presentation.
Dickinson et al. categorized (5) 303 patients based on achieved surgical margins: contaminated, <1mm but clear, 1-4mm, 5-9mm and 10-19mm. They concluded that “the margin can safely be as narrow as 1mm in terms of low local recurrence rates”. However, their study did not consider critical factors such as tumor size and resection difficulty and included previously treated patients. Bilgeri et al. categorized all R0-patients, including those who presented with LR and DM, into four categories: <1mm, 1-5mm, 5-10mm and > 10mm and found that wider margins were associated with better outcomes up to 5mm, beyond which there was no significant improvement in LRFS and overall survival (10).
In a broader pooled study population including patients with both low- and high-grade STS, McKee et al. demonstrated that those with microscopically positive margins and close negative margins (1–9 mm) were at increased risk for LR and distant metastases compared to patients with clear margins measuring ≥ 10 mm. Having said this, the study is further limited by its low percentage of adjuvant radiotherapy (38%) and its inclusion of re-resections.(6) In the most selective study Fujiwara et al. were able to show that in a cohort composed only of patients with infiltrative STS, a surgical margin of greater than 10mm was associated with an improved LR-free survival.(9) Albeit these findings are limited because not all negative margins could retrospectively be sorted into the right category.(10) In a prospective study including all types of STS and previously treated patients, Sampo et al.(12) reported that the LRFS correlated with increasing surgical margin as far as 4cm and indicated that a surgical margin of 2-3cm provided “reasonable local control”.
Consistent with previous research, our findings demonstrated that negative margins are significantly superior to positive margins, with a hazard ratio (HR) of nearly 8 (HR 7.845; CI 3.984–14.450). Adjuvant radiotherapy also played an important role in achieving local control, reducing the HR of local recurrence (LR) by more than threefold (Table 3). Additionally, our study showed a direct correlation between local recurrence-free survival (LRFS) and margin width (Figs. 1 and 2): larger surrounding non-reactive tissue in the tumor specimen was associated with longer LRFS. Radio- and chemotherapy, as well as tumor stage, were relatively evenly distributed among the subgroups. Patients with resection margins > 5mm had the best 5-year LRFS at 93%, followed by those with margins of > 1-5mm at 76%. Our 5-year LRFS is slightly lower than most reports in the literature but aligns closely with studies focusing exclusively on high-grade soft tissue sarcomas (STS), particularly those examining wide resection margins (10,29). Notably, in our study, no LR occurred when the margin was > 5mm and adjuvant radiotherapy was applied, a result significantly superior to patients with margins < 5mm and adjuvant radiotherapy (p ≤ 0.023).
Age at surgery attained statistical significance in both uni- and multivariate analysis concerning LR, although it only slightly increased the HR. Duration of surgery, considered a composite measure of tumor characteristics such as size, depth, and stage, was significant in bivariate analysis, not aligning with expectations since individual tumor characteristics did not independently reach significance.
It becomes apparent that there is more to predict LRFS than only metric treatment options. Proximity to vascular bundles, nerves or fascia has become a more critical decision-factor in recent years. Kawaguchi et al.(7) first suggested that certain barriers such as the above have a stronger resistance against STS infiltration than normal tissue. O’Donnell et al.(30) emphasized the impact of different clinical settings in patients with positive margins on LRFS and supported the practice of close dissection for the preservation of critical structures even when this leads to microscopically positive margins. They hypothesized that the infiltration of critical structures can hint to a more aggressive nature of the tumor in general. However, this must be weighed against the potential risk of proposing an untrue or falsely assessed surgical barrier which led to the worst 5-year LRFS in their study. Our findings support that the question of what constitutes adequacy should be considered in context but that there is a good chance of achieving local control when a resection margin of more than 5mm can be achieved, particularly when coupled with adjuvant radiotherapy, which is standard practice in most cases.
Table 6
Overview over the literature: resection margins and their impact on LR in STS patients
| Reference | Nr. of patients | Resection margin categories | Impact on LR | Limitations |
| Dickinson I.C. et al., ANZ J Surgery 2006 | 303 | Contaminated, < 1, 1–4, 5–9 and 10-19mm | margin can safely be as narrow as 1mm | Including re-resections, no respect to tumor size and difficulty of resection |
| Bilgeri A. et al., Cancers (Basel). 2020 | 305 | Contaminated, < 1, 1–5, > 5 and > 10mm | A margin of > 5mm is sufficient, wider margins do not benefit the patients | Including re-resections; margin, tumor size and age are linked to RTx and CTx. |
| McKee MD et al., J Surg Oncol. 2004 | 111 | Contaminated, 1–9 and > 10mm | Margins > 10mm are optimal for extremity resections | Including low-grade STS and re-resections, low percentage of adj. RTx and CTx |
| Fujiwara T et al., Eur J Surg Oncol. 2020 | 305 | Contaminated, 0.1–0.9, 1.0-1.9, 2.0-4.9, 5.0-9.9 and > 10mm | A margin of > 10mm is advocated | not all negative margins could retrospectively be sorted into the right WHO-category |
| Kainhofer V et al., Eur J Surg Oncol. 2016 | 265 | UICC- and R-Classification | R0 resections are superior when classified according to the UICC-classification | Including low and high-grade STS and re-resections; treatment for atypical liposarcomas changed during the second half |
| Gundle KR et al., J Clin Oncol. 2018 | 2217 | R-Classification, R + 1-Classification and TMCC classification | An R + 1mm classification reduced LR-differences between R1 and R0, but the R-classification best determined the risk of LR | Single-center study, treatment protocol has changed over the years; tumor sampling errors cannot be ruled out |
| Sampo M et al., Br J Surg. 2008 | 270 | < 0.4, 0.4-2.0, > 2.0mm | A surgical margin of 2-3cm provided reasonable local control, even without the use of radiotherapy | Including low and high-grade STS as well as post radiation STS and patients who received amputation. |
| Our findings | 207 | Contaminated, 0.1-1, > 1–5 and > 5mm | A margin of > 5mm is advised | Anatomical boundaries could not be evaluated, treatment protocol has shifted over the years |
R vs. UICC-Classification
In 2016, Kainhofer et al. (4) initiated a discourse regarding the prognostic significance of various classifications of resection widths in patients with STS. They primarily focused on the R-classification as suggested by the American Joint Committee on Cancer (AJCC)(22) and the R + 1 classification suggested by the International Union against Cancer (UICC)(31). Their results favored the UICC-classification. However, two years later, Gundle et al. (14) conducted a comparative assessment of both systems and determined that the R-classification was more adept at assessing the risk of local recurrence (LR) within a competitive risk framework. Our findings cautiously align with this perspective, as we observed no enhancement in LRFS with the R + 1 (UICC) classification when juxtaposed with the traditional R-classification (Fig. 5).
While study results are relatively consistent on how resection width and margin status relate to LRFS, an association with DMFS remains questionable (13,19).
It must be said that neither the present, nor one of the studies mentioned above demonstrate any statistically significant correlation between the quantitative resection margin and metastasis free survival, underscoring the necessity for further research into this topic. While this study could show that margin width and subsequently margin status were independent risk factors in univariate analysis, their significance waned in a multivariate setting. In contrast, adjuvant chemotherapy and LR had a decisive impact on DMFS: as shown in the LRFS-analyses, there is an inverse relationship between LR and resection width. Analyzing the present study, LR itself proves to be a significant predictor for DM but the causal association has been described as weak by Trovik et al.(13) before. The authors argue that since inadequate margins were the highest risk factor for LR, inadequate margins should also be a significant prognostic factor for DM in an analysis not including LR. However, the authors did not show this in their analysis. Upon removing LR from Cox regression analysis in the present study, margin status assumed critical importance (p = 0.021), highlighting its role in influencing DMFS. Radiotherapy, while significant for LR, did not retain significance for DMFS, suggesting a nuanced interplay between treatment modalities and metastatic behavior in STS. Thus, the present study underlines that negative margins are vital in the prevention of both LR and DM Moreover, larger-scale studies investigating the interaction between surgical resection margins, (neo-)adjuvant therapy, and the biological properties of STS are imperative for devising effective treatment strategies aimed at prolonging DMFS
Similarly, the impact of margin status on OS remains discussible. While a meta-analysis by Jang et al (11) and a recent study by Bilgeri et al.(10) concluded that margin status does affect OS, Jang et al. acknowledged several studies that showed contradictory results.(19,33–35) However, they acknowledged the low number of participants in the mentioned studies. The latest by Chouliaras et al.(8) could not find any correlation either. In another approach, Willeumier et al. argued that due to the aggressive nature of high-grade STS and its decrease in the 5-year survival rate it is difficult to determine the effect of surgical margin on survival(29) and that the effect will manifest over a longer time on patients who escape early DM(11,26), which is why we set the endpoint for OS at 10 years. While margin status could not uphold its significance in a multivariate setting, the major prognostic factors for OS were LR and DM with a HR of 2,8 and 8,7, respectively. Given the pivotal role of both local recurrence (LR) and distant metastases (DM) as primary prognostic factors for overall survival (OS), it follows that margin status should significantly influence OS as well. Upon excluding LR and DM in a Cox-regression analysis with respect to OS, we observed that R1-resection emerged as a negative prognostic factor (p = 0.045, HR 2.209). This underscored the importance of achieving negative margins to optimize OS Additionally, the findings in the present study suggest that adjuvant radiotherapy and chemotherapy can improve OS, highlighting the potential benefits of adjunctive therapies in enhancing survival outcomes. While specific details on adjuvant radiotherapy and chemotherapy were not collected in this study, optimizing the administration of these therapies akin to Schliemann et al(36) in R0-resected high-grade STS patients is important.
Interestingly, OS was influenced by wound-healing complications. However, these were mostly noted in patients with positive margins (42%) whereas in other subgroups, only around a fifth of the patients developed wound-healing complications (19%, 21%, 26%). Interestingly, we observed that OS was influenced by wound-healing complications, with a higher incidence noted in patients with positive margins (42%) compared to other subgroups (19%, 21%, 26%). This disparity may be attributed to several factors, including the advanced age of patients with positive margins, their tendency to receive less adjuvant therapy, and the larger average tumor size among patients with wound-healing complications as the tumor might have been in proximity of critical structures.
Limitations of our study
We acknowledge the following limitations in our study. There was an uneven distribution of patients receiving neo- or adjuvant chemotherapy within the subgroups, and the overall numbers were small. Additionally, exact histopathological resection margins were untraceable in 26 patients, limiting our analysis to 177 patients when grouped by resection width. However, when we analyzed by resection status, all 203 patients could be included. To minimize distortions, we established strict inclusion criteria. However, our cohort still included patients with various soft tissue sarcoma (STS) entities, and we could not account for anatomical barriers beyond the superficial or subfascial location of tumors.