Desmoid tumors (DTs), also known as aggressive fibromatosis and desmoid-type fibromatosis, have an incidence of approximately 5 per 1 million persons and are characterized by no metastasis, but local progression and recurrence can occur (1, 2). In general, DTs can be divided into sporadic and familial adenomatous polyposis (FAP)-related DTs. Sporadic DTs are associated with a mutation in the CTNNB1 gene and are usually located in the abdominal wall or outside the abdomen (3). DTs on the abdominal wall are most common in women within one year after pregnancy or childbirth. FAP-related DTs, which are associated with mutations in the APC gene, usually occur in the intra-abdominal region, including the mesentery, retroperitoneal area, or pelvis, and can be very large (2, 4).

DTs can occur at any age; the typical age of onset is 15–60 years, with a peak of onset at approximately 30 years (5, 6). The pathogenesis of DTs is related to many factors, including trauma, hormones, and heredity. Trauma, including surgery, may cause DTs (7). It was found that the incidence was higher in females, a significant number of women had a recent pregnancy history at the time of diagnosis, and the risk of progression during pregnancy was higher, which suggests that the disease is related to hormones (8). Heredity mainly refers to FAP patients being prone to developing DTs. It has been reported that patients with FAP are 1,000 times more likely to develop DTs than those without FAP (7). In addition, 21%–31% of FAP patients may develop DTs (6, 7). The natural history and clinical symptoms of desmoid-type fibromatosis vary significantly. In some cases, the tumor may be stable for a long time without obvious clinical symptoms, while in other cases, it may rapidly increase in size. Intra-abdominal DTs can cause intestinal obstruction (9). DTs located in the extremities are generally not life-threatening conditions that would cause a patient’s death, but they can cause patients to have pain, restricted mobility, and loss of function (9). In some studies, one of the problems faced by patients with extremity tumors is poor prognosis; that is, the probability of recurrence after surgical resection is higher than in other locations, and some patients may even face the risk of amputation (10, 11). Therefore, care must be taken in the management of patients with extremity tumors.

With an improved understanding of the natural course of DTs, changes in the treatment of DTs have also emerged. According to the latest treatment guidelines, the preferred treatment for DTs is no longer surgery or other treatments but close conservative observation (12, 13). There are a number of patients whose tumors do not progress during observation (14). However, surgery remains an important treatment option when patients experience pain or pressure or other clinical symptoms, and chemotherapy, hormone therapy, and targeted therapy may also be considered when the tumor cannot be resected or the surgery would lead to serious functional loss (15, 16). Radiotherapy can also treat DT. James et al. suggested that radiotherapy is recommended for unresectable or recurrent DT (17). According to the latest NCCN guidelines (2020.V2), radiotherapy is not recommended for intra-abdominal DT. For patients with superficial trunk or extremity DT, definitive radiotherapy can be used as first-line or second-line treatment. On the other hand, radiotherapy may also induce secondary malignant tumors. Currently, one of the major challenges in surgical treatment is the high postoperative recurrence rate, which in some studies is as high as 60% (18). To date, recognized risk factors for recurrence include age, site, tumor size, and genetic mutations. Patients with positive expression of Ki-67 and a high signal on T2-weighted MRI have a higher probability of recurrence after surgery (19). In addition, higher PARP-1 (a DNA repairing enzyme) expression is also associated with earlier recurrence of DTs (20). In 2013, Crago et al. constructed a nomogram, which included age, location and tumor diameter based on these recurrence factors, that can roughly predict the probability of recurrence in DT patients. In addition, whether postoperative recurrence of DTs is related to other factors remains to be explored. For example, tumor number is rarely mentioned in previous studies because of the rarity of multiple tumors. In addition, the development of a tool that allows both doctors and patients to easily predict the probability of postoperative recurrence and to provide patients with a more reasonable treatment plan are also promising research directions. In the present study, we constructed a new web-based nomogram that shows good discrimination and calibration in both the development population and validation population and provides prediction probabilities that are more accurate, convenient and visual.

As a borderline tumor, although the mortality rate of DTs is very low, their unpredictable natural history still brings challenges to clinical treatment. The Sarcoma PAtients EuroNet (SPAEN) and European Organization for Research and Treatment of Cancer (EORTC)/Soft Tissue and Bone Sarcoma Group (STBSG) recommend active surveillance as the preferred initial treatment for DTs (26). When patients have symptoms of compression, surgery is still an important treatment, but the high recurrence rate after surgery remains a major problem for patients and doctors. Therefore, we developed and validated a nomogram and further visually integrated it into a web version for the first time. Clinicians can accurately predict the recurrence rate by selecting several parameters on the web page to strengthen postoperative monitoring and provide clinical guidance.

In contrast to the nomogram created by Crago et al. in 2013, which includes age, diameter, and location, our results show that although location was associated with recurrence in the univariate Cox analysis, the recurrence rate of tumors located in the extremity was higher than that of body tumors (p = 0.033), but location was not an independent predictor of recurrence in the multivariate Cox analysis (p = 0.423) (10). Several other studies have also found similar results in univariate and multivariate Cox analyses (23, 24). However, in two other large cohort studies, even in the multivariate Cox analysis, the conclusion that location is related to recurrence was still obtained (21, 22). Our next work will continue to collect medical records on patients with DTs in the next few years and re-analyze the relationship between tumor site and recurrence when the number of patients has further increased.

In addition, compared with the nomogram created by Crago et al. in 2013, we believe our web-based nomogram has two innovations as follows. First, we added the tumor number as a predictor for the first time, and indicated that the recurrence rate of patients with multiple tumors was higher than that of patients with single tumors (p = 0.001). Actually, multiple DTs usually occur in FAP-related patients but are rare in sporadic DTs (2). At present, the proportion of patients with multiple DTs reported in most of the literature is less than 10%. James et al. found 5 patients with multiple DTs among 101 patients, accounting for 4.95% (17). Marco et al. found 6 patients with multiple DTs out of 92 patients, accounting for 6.5% (8). Huang et al. found 14 patients with multiple DTs among 151 patients, accounting for 9.3% (27). Salas et al. found 10 patients with multiple DTs among 426 patients, accounting for 2.3% (21). Among them, only Huang et al. discussed the relationship between the tumor number and postoperative recurrence and indicated that the tumor number was not related to postoperative recurrence (p = 0.422). Two other studies with more than 400 patients also did not study a relationship between tumor number and recurrence (10, 22). We speculate that the relationship between tumor number and recurrence has rarely been studied before because multiple DTs are rare. In our study, the proportion of patients with multiple DTs was only 5.2%, which was approximately the same as that reported in other literature. Our next work is to cooperate with other hospitals to collect data from patients with multiple DTs to further confirm our conclusion. Given the rarity of multiple DTs, the association between tumor number and recurrence needs to be further confirmed in other large studies. Second, in terms of the nomogram constructed by Crago et al., patients and doctors need to find the corresponding score of each variable and add these scores to calculate a total score. Then, the probability of recurrence is predicted based on this total score. However, the steps are relatively cumbersome, and the result is only an approximate probability, without an overall understanding of the prognosis. The web-based nomogram we constructed is more visual than the traditional nomogram, eliminating the cumbersome steps of calculating scores. The web-based nomogram can directly display an accurate recurrence probability after selecting the corresponding clinical parameters and follow-up time directly on the web page and can also display a 95% CI. On the other hand, after selecting the corresponding parameters, the RFS curve can also be displayed, giving a clear understanding of the overall prognosis of the patient. Therefore, this web-based nomogram is highly visualized, accurate and intuitive.

In this study, adjuvant radiotherapy did not reduce the postoperative recurrence rate. In fact, the role of postoperative adjuvant radiotherapy has been controversial in the past few decades. Different studies have come to different conclusions (28–31). A meta-analysis has indicated that adjuvant radiotherapy can reduce the recurrence rate of patients with positive resection margins, but for patients with negative resection margins, adjuvant radiotherapy cannot reduce recurrence or improve the local control rate (32). This may also be the reason why we did not obtain the conclusion that adjuvant radiotherapy reduces recurrence in our study, as all of these patients had a negative margin status (R0) in our study. For the Chinese population, most studies still report that postoperative adjuvant radiotherapy cannot reduce recurrence (33, 34). Among them, the study of Huang et al. included 214 patients (24). Therefore, the significance of adjuvant radiotherapy remains to be further studied.

In terms of clinical application, we believe that the main application value of this web-based nomogram lies in the decision-making before surgery and enhancement of postoperative surveillance. First, for patients with a relatively high recurrence rate, such as patients aged <=30 years with a tumor diameter >10 cm and multiple tumors, the web-based nomogram indicated that the 5-year RFS rate was only 16%, suggesting that such patients are extremely prone to postoperative recurrence. In this case, surgery is not the preferred treatment, and doctors or patients can choose other treatments such as radiotherapy alone or systemic drug treatment. For patients with a low recurrence rate, such as those aged >30 years old with a tumor diameter <=5 cm and a single tumor, the web-based nomogram indicated that the 5-year RFS rate was 92%, which indicates that the probability of postoperative recurrence is low, and surgical treatment has shown its superiority compared with other treatments. Intensive follow-up is all that is needed after surgery. It is important to point out that this is a nomogram that has only been tested on patients undergoing surgery, and factors related to recurrence existed before surgery. Through this nomogram, the effectiveness and success rate of the operation can be predicted. For patients with a high recurrence rate after surgery, other treatments can be chosen. Therefore, the greater clinical significance of this nomogram lies in predicting the efficacy of surgery and the choice of treatment methods. Second, this web-based nomogram will also help enhance postoperative surveillance. For patients who have received surgical treatment, this web-based nomogram can accurately predict the recurrence probability at each time point after surgery so that doctors and patients can have a clearer understanding of the prognosis and strengthen postoperative surveillance. According to the latest Desmoid Tumor Management Guidelines published by “The Desmoid Tumor Working Group” and NCCN (Soft tissue sarcoma, 2020, V2) guidelines, DT can be followed by MRI or CT, especially MR I (12). Postoperative MRI images of recurrent DT mainly showed enhancement of T2-weighted, T2-FS and STIR sequence signals, accompanied by the enhancement of nodular or mass‐like T1-weighted postcontrast sequence signals (35, 36).

The limitations of this study cannot be ignored. First, the type of CTNNB1 gene mutation was not included in our study, and several previous studies have confirmed that the type of CTNNB1 gene mutation is also associated with the recurrence of desmoid fibromatosis. There are three types of mutations in the CTNNB1 gene, namely, T41A, S45F, and S45P, among which patients with the S45F mutation have the highest recurrence rate (1, 37, 38). We have tried to find the gross specimens of these patients from our hospital’s tissue sample bank to detect the mutation type of the CTNNB1 gene. Unfortunately, despite our best efforts, we were able to find only a few samples, which were insufficient to study the effect of the mutation type on recurrence. In our ongoing work, we will focus on collecting the gross specimens of patients to detect the mutation type of the CTNNB1 gene in patients and predict RFS more accurately. Second, 80 patients were lost to follow-up out of a total of 385 reviewed patients, and these lost cases may bias the final outcome. Third, this study is a retrospective study, which may produce bias. A prospective investigation would better enhance and validate these results. Fourth, we could not find an available non-Chinese patient cohort to validate the nomogram, even though we validated it on a dataset from another Chinese cancer center. Nevertheless, our findings could still benefit clinical practice because the web-based nomogram can strengthen postoperative monitoring.

In summary, we developed, validated, and visualized a web-based nomogram including age, tumor diameter, and tumor number based on independent risk factors associated with recurrence to predict RFS in patients with DTs. In the future, the use of multiple technologies to further search for biological or clinical markers that predict the recurrence of DTs is a development direction of surgery, such as the combination of radiomics and artificial intelligence to find the imaging characteristics of DT recurrence. The preoperative evaluation of these characteristics can further predict the risk of recurrence to determine the optimal treatment plan.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Tianjin Medical University Cancer Institute & Hospital ethics committee and performed according to the Helsinki declaration. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

HL drafted the manuscript. KH and YC provided part of data and revised the manuscript. TL, TY, ZL, CZ, LX, YC, and JY revised the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the Key Nature Science Foundation of Tianjin [18YFZCSY00550 to JY].

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.