Desmoid tumours are rare and have a well-recognised unpredictable natural history. To date, there remains insufficient and conflicting evidence to define any predictors of this varied behaviour. This translates into a persistent challenge in selecting the best treatment option for an individual patient. We, therefore, sought to evaluate various clinicopathological attributes in a series of primary sporadic desmoid patients managed at a single large soft tissue sarcoma unit.
In this study, we have excluded patients with hereditary fibromatosis, which is known to have an aggressive clinical course18. Also, to reduce the heterogeneity of the group we have not included patients who presented with recurrent disease.
In recognition of the fact that a good proportion of desmoid tumours can regress spontaneously, an upfront wait and watch strategy is becoming the consensus first-line management approach12,13,19. In our study, 69.5% of patients were offered an initial trial of active surveillance. The rate of regression was 24.2%, while the rate of progression and 5-year progression-free survival was 28.8% and 87.8%, respectively. These results are consistent with and rather better than the internationally documented figures20,21. Large tumour size on uni and multivariate regression analysis was found to be a significant factor with the tendency to progress over time. However, none of the other studied parameters, including association with pregnancy, were found to correlate on multivariate analysis. A recent study of 168 sporadic primary desmoids of all anatomical locations from a national high-volume centre showed a radiologic progression rate of 36% on active surveillance. The regression rate and rate of stable disease in their study was 27% and 36% respectively21. This mirrors the results of our study as the population cohort is essentially similar other than the inclusion of limb fibromatosis cases in the French series. They also performed a regression analysis to find age younger than 50 years to be a significant contributor towards a higher progression rate (p=0.046), which remain an insignificant contributor in our study group.
In the series by Van Houdt et al., treatment was offered to 46% (limb and trunk) and 43% of patients in our data set20. The progression-free survival at 2-years (85.7%) was also comparable. The indication for active treatment in our study was morphologic progression with or without symptoms in 67.9% and symptomatic disease alone in 32.1%, which remains consistent with their results of 63% and 32%, respectively.
A systemic review of 211 patients from five centres reported an overall rate of progressive disease ranging from 4-34%, with a median time to progress between 14 and 20 months and time to regression of between 6 and 130 months. However, all progression events were recorded in the first two years after diagnosis. The studies included in this review have given widely varying results owing to the heterogeneous patient populations including both primary and recurrent disease as well as sporadic and hereditary fibromatosis. The study also included patients who had had previous surgery, a factor associated with increased risk of progression22. Mean time to progress in our study (14+2 months) was nevertheless consistent with the above results. As symptomatic disease is the cause for a failed observation approach in about a third of patients, advancement in medical therapies for symptom control along with better psychological support to help patients could pave the way to superior outcomes7,8,20,23.
Several studies to date have analysed tumour recurrence after surgical resection. In a systemic review of 781 cases from 16 papers/studies, Smith et al., reported a local control rate of 42 to 86% for patients managed with surgery alone and a rate of 69 to 84% with the multimodality treatment. Their reported recurrence rate varied from 14 to 58% in included individual series22. Recurrence rate in our study after resection of primary disease was 28.6%, with a two-year recurrence-free survival of 77.1 % and a median time to relapse of 17 months. A large single centre series of 495 surgically resected patients by Crago et al. reported a recurrence rate of 23% and 5-year recurrence-free survival of 69%24. On multivariate analysis, they found age (less than 26 years), large tumour size and tumour location (extremity and chest wall) associated with a shorter RFS. In another relatively recent series of 177 operated patients by Muller et al.25, the observed recurrence rate over a median follow-up interval of 40 months was 29.4%, with a five-year RFS of 61%. The study populations in both of the above series were fairly heterogeneous. Both datasets included hereditary fibromatosis, recurrent disease (23 and 25% of the total population in each series, respectively), adjuvant radiation treatment (18.7% and 20 %), along the addition of systemic therapy in an additional addition subset in each. Nevertheless, summarising the above recent data with our results, documented recurrence rates remain less than 30% irrespective of the use of adjuvant or neo-adjuvant treatments. Microscopically involved tumour margin status in our surgically resected group failed to yield any correlation with the outcomes on both uni and multivariate analysis. Mullen et al., in their series found negative margin status (R0 resection), an independent prognostic factor with a better event-free survival on multivariate analysis. Nevertheless, the results may have been influenced by the addition of adjuvant treatment (29%), specifically radiation therapy in 20%. The recurrence rate for primary tumours subset in the study was 32%, with a median time to relapse of 14 months, results which are comparable to our study group25. Margin status (R1 vs R0) was not associated with altered outcome in the study cohort by Crago et al., but an R0 resection in the subgroup with smaller tumours <5cm was associated with a longer RFS (p=0.007)24. Systemic review by Smith et al., has reported the R0 margin resection rate to a range between 18% to 100% 22. Overall, margin status has been variably or poorly defined and remains a debatable predictor for determining recurrence. However, where possible, an R0 resection (margin >1mm), or at least a negative margin, is desirable22,24,25,26,27.
The overall adverse event-free survival has been described as better in the conservatively managed than the upfront surgery group (58 vs. 53% (p=0.415)) 21. Similar findings were noted in our study group with a statistically significant difference on Cox regression analysis for overall adverse event-free survival. Overall rate of event-free survival over two years in our study was 83.8%, which is much higher than 56% in the study by Penel et al21. Only a few cases had a relentless course in present study (5.3%) and required complex treatment strategies. That included, in addition to surgical resection, multiple chemotherapy courses, cryoablation in 2 patients, radiation therapy and RFA to one each.
As the management approach towards desmoids has evolved during the span of this study, more patients in the latter part of the study were offered active surveillance as a first-line treatment option. Also, the national guidelines for sarcoma patients have recently led to better referral pathways with more patients presenting upfront rather than after excision biopsy of these tumours in a local hospital12. Considering the above-mentioned selection biases and the inherent biases with a smaller subgroup population, we have not drawn a direct comparison analysis between the two management groups. However, the local control rate in both groups of primary sporadic truncal desmoids compared favourably (28.8% in W&W or 28.6% in the selective surgical group) with the previously reported series20,21. Since the local control rate is very similar, an active surveillance approach in asymptomatic patients should be considered first-line, with patients that progress or become symptomatic being offered either medical management, or surgical resection aimed at achieving an R0 resection and preserving function 7,11,12.
The results from various single centre series have failed consistently to correlate disease course and clinicopathological features. A collaborative effort with a larger homogenous data set to elucidate the contribution of these on the clinical behaviour of this rare pathology and its management is imperative. Moreover, although we included beta-catenin staining on immunohistochemistry in our study, results of mutational analysis of the CTNNB1 gene were not studied, due to its relatively recent inclusion in clinical practice. Additional future perspectives also lie in identifying different molecular mechanisms to correlate with tumour biology and tumour behaviour in fibromatosis patients, including the role of different CTNNB1 subtypes. This may pave the way to developing stratified, and targeted treatment approaches towards tumours with potentially more aggressive behaviour14,27,28.