Recurrent RPS disease after curative treatment is associated with a worse prognosis. These patients offer several challenges for surgery, because of the extension or location of the recurrence, but also for the patients' performance status. Furthermore, nowadays the availability of newer medical therapies requires the attentive selection of patients as candidate for surgery, medical therapy or integrated treatments. Some tumoral characteristics determine the pattern of recurrence, as well as the median time of relapse: WD-LPS generally has local recurrences with an indolent behavior over time, whereas LMS tends to have a systemic spread, both early and late in the course of disease .
While preliminary works have questioned the role of surgery in locoregional recurrence , larger cohort studies  have instead demonstrated a significant association between resection and survival in this sub-group of patients. Similarly, Lehnert et al  have shown that patients who present with local recurrence from RPS have an increased risk of experiencing further local tumor recurrence, but with a consequent reoperation plan, a respectable long-term survival was achieved. Similarly, in our series LR patients experienced further recurrences, in particular in the original site (81.8% of total). At our centers, a relatively aggressive approach was taken to resect LR by performing a wider excision as it might have been done for primary RPS. This was in part related to the fact that most primary tumors had been removed in peripheral hospitals, sometimes without compartmental surgery, and only complex cases or recurrences were then referred to our tertiary care institutions. So, our data are in favor of a surgical wide treatment that should be considered when an oncological safe resection could be obtained, even if percutaneous locoregional treatments have been taken into consideration in literature in specifical settings, such as in patients initially unfit for general anesthesia, or to treat single nodules in the retroperitoneal space, such lumbar muscles or iliopsoas one, or parenchymatous organs such as liver . These treatments, were chosen in our series prevalently for further recurrences or combined with surgical resection for the first recurrence.
The management of systemic recurrence is more complex. Although published data are largely retrospective and included heterogeneous populations, there is a consensus to support metachronous metastasectomy . The proper treatment depends on the extension of the disease and should be tailored in a multidisciplinary manner. In our experience, we reported no difference in terms of DFS and OS after surgical resection in the LR versus the DM group with a mean OS of 61.99 and 56.83 months, respectively. This might be related to the fact that up to 66% of the total, distant metastases were usually of limited extension in our series and were treated with parenchymal sparing liver or lung resection (that also can explain the lower operative time for the LR and LR + DM group). The vast majority of patients received CT and three patients, two with LMS and one with DD-LPS, are still alive at a follow up of 84, 62 and 20 months. Indeed, for this specific sub-group, recent clinical literature and systematic review describe long survival, especially when affecting the liver . For these reasons, even if these are metastatic lesions, in selected cases, such as those with an interval between primary surgery and recurrence > 6 months, and small liver lesions, it seems reasonable to take into consideration their surgical resection [14, 15]. Although the literature search did not identify any studies comparing surgery versus chemotherapy in the DM group, a recent study by Smolle et al.  compared surgical vs. non-surgical management in a more general cohort of patients with sarcoma lung metastases. This study demonstrated surgery to be associated with significantly longer OS (10-year OS: 23% vs. 4%, p < 0.001) and this advantage are independent of baseline characteristics. Moreover, in our opinion, a radical approach should be taken into consideration also in the case of resectable DM after previous surgery for recurrent disease , because it could influence positively the OS, as we reported from our experience in which repeated interventional procedures is the leading factor both at univariate and multivariate analysis.
The present work describes one of the highest percentages of repeated surgery for RPS recurrence, as we described "re-do" surgery in 50% of patients who presented new resectable recurrences after the surgical treatment of the first RPS recurrence, with a percentage of 63% in LR group (also in case of distant spread in one patient), and 28% in DM group. Instead, for the LR + DM group albeit all cases experience new recurrence, no one was reoperated. This is in line with the trans-Atlantic RPS Working group consensus document  that state that surgery has a more limited role in patients with liver disease associated with an extra-liver disease or multifocal abdominal metastases. However, all these patients were operated on for the first recurrence with complete resection of all nodules/metastasectomy. To date, very scarce literature is present on the surgical treatment of both local and metastatic RPS recurrences, in particular with multiple peritoneal implants or systemic spread. This aspect could be considered another key point of our analysis, although on a limited number of patients, and it seems to support the possibility of a surgical approach in case of RPS recurrences, also in patients with LR + DR, as we obtained satisfactory short-term and long-term results compared to other studies on this argument .
In the LR + DR group, 33.3% of patients had positive margins at the final histological report, but this may be due to the fact the large masses had been removed. Margin resection is a risk factor that increased mortality in univariate analysis, but not in multivariate one. However, the question of positive margins is still very debated. Even if different investigators found no benefit for tumor incomplete resection compared to exploration alone , a recent study addressed the question of clinical outcome in patients with retroperitoneal LPS, focusing on positive margins and surgical debulking . These authors found a survival benefit in patients after incomplete resection with a median survival of 26 months compared to exploration alone with a median survival of 4 months. A meta-analysis of patients undergoing incomplete resection of RPS found an improvement in survival time after cytoreductive tumor resection . We can then say that even if a good pre-operative evaluation is mandatory to offer surgical resection to those patients who are very likely to achieve an R0-resection, in those cases who achieved an R1 resection there might still be some benefit deriving from surgery, even if in these cases the integration with radiotherapy or medical therapy becomes fundamental.
Our study has some limitations due to its retrospective nature. First of all, a possible selection bias could be present. An unknown percentage of patients with resectable recurrences was probably treated only by the oncologist without a surgical evaluation. Moreover, we included patients enrolled in a long period in which medical therapies have undergone important improvements and nowadays we have various valid treatment options with different impacts in terms of cancer-specific and overall survival. Furthermore, although all patients have been always followed by oncologists after surgery (even in different centers and regions), we were not able to collect detailed information about the medical therapy regimen they underwent after the surgical operation, and therefore we were not able to report detailed information about these data. Furthermore, local treatment such as percutaneous thermal ablation might have a role but no randomized studies compare the efficacy of surgical procedures versus the only medical therapy versus these treatments. Concerning systemic therapy, there is not enough evidence to use post-operative therapy after surgical recurrence resection . Medical therapy, on the other hand, could play a role before interventional procedures to downsize the recurrence, especially locoregional ones. Finally, the small sample size, particularly in the DM group and LR + DM group, limits a reliable statistical analysis.
To conclude, data emerged from the present series suggested additional information to the RPS recurrence scenario and should encourage a thorough multidisciplinary evaluation of patients with recurrent RPS performed by general surgeons, interventional radiologists, and oncologists to take into consideration a possible surgical treatment of the RPS recurrences, also in patients who have already undergone surgical resection of RPS recurrence and developed new treatable RPS recurrences during the follow-up. Although no differences in terms of DFS and OS between our three groups were detected, the importance of repeated interventional procedures seems to be particularly important in the LR group and DM group. Locally directed therapy, such as surgical and ablative approaches, may be integrated and can offer a potentially long survival in selected patients. Therefore, even in those cases in which RPS recurrence is located in the liver or lung, when an R0 resection could be obtainable, surgical resection could be taken into consideration, especially in high volume surgical centers, because surgical resection has been demonstrated to have a positive impact on long-term oncological results. LR + DM group is the worst survival group in which no further interventional maneuvers were taken into consideration as the progression of intra-abdominal disease; however, the DFS and OS seem to be reliable compared to other works in which the LR + DM relapse was not removed supporting the importance of surgery also in this setting with related low morbidity. Probably, a combination of therapies and progress in the management will be crucial for this kind of patient in the future. However, given the rarity of this pathology and particularly for the resectable RPS metastases, this arena would benefit from prospective multi-institutional collaborative efforts to define the role of peri-operative chemotherapy as adjuncts to surgical excision and percutaneous ablation and to delineate the optimum sequences of these therapies in the treatment of these patients as well as a detailed risk evaluation for survival.