The benefit of early detection of local recurrence depends on the availability of therapeutic options that can prolong survival. Radical compartmental resection with or without adjuvant radiotherapy and/or chemotherapy may provide long-term salvage in patients with a local recurrence of soft tissue sarcoma [34]. The risk for local recurrence depends on tumor site, size, grade, and adequacy of surgical margins. Although about 2–20 % of all resections will have positive margins and thus increased risk for local recurrence, its impact on overall survival remains controversial [1, 4, 35].

Several guidelines have been recommended for the follow-up of soft tissue sarcoma consisting of a combination of clinical history, physical examination, blood tests, chest radiographs, CT, and MR imaging [36–42]. The ACR advises MR follow-up at least twice a year during the first 5 years after treatment of the primary tumor and once a year 5–10 years after treatment. The ESSR and ESMO advocate 3–4 MRs during the first 2–3 years in intermediate- or high-grade sarcomas and 2–3 MRs during the first 3 years in low-grade sarcomas. An MR once a year 5–10 years after primary treatment is also advocated by ESSR [34, 43–47]. Most institutions rely on consensus-based guidelines due to the absence of evidence-based guidelines. Surveillance strategies that, through early detection and treatment, improve survival and quality of life while minimizing costs have yet to be identified in randomized clinical trials. Only few studies have been reported on the efficacy of surveillance strategies for the follow-up of soft tissue sarcoma [37, 38, 48]. According to Whooley et al., clinical assessment and physical examination are the most useful tools for evaluating locoregional recurrence, whereas routine MR imaging of the primary tumor site and laboratory blood tests appear ineffective strategies. In a retrospective review of 141 patients, they detected by routine annual imaging only one asymptomatic local recurrence; all others were found on physical examination of the primary site [37, 48]. However, MR imaging has shown to be useful in patients in whom physical examination is hampered due to radiotherapy changes.

When indicated, MR imaging is the most useful technique for identifying suspected local recurrence or residual disease after incomplete resection [34, 49]. An MR using a T2-weighted sequence with frequency-selective fat saturation or short tau inversion recovery (STIR) or Dixon sequence with water reconstruction is considered to be the most useful first step for detecting recurrent tumor (Fig. 26.3). The morphology of the lesion and the signal intensity contribute to the definition of its character. Low signal intensity on T2-weighted images or diffuse high signal intensity on T2-weighted images excludes tumor recurrence in 99 % of patients. Mature scar tissue usually exhibits low signal intensity (Fig. 26.4) because of its fibrous tissue content, as described in previous studies. Diffuse high signal intensity with a feather-like appearance, without mass effect, generally represents post-therapy change or inflammation. High signal intensity mass-like lesions on T2-weighted images require further examination with intravenous Gd-chelates [50].

Fat-suppressed T1-weighted fast spin echo sequences after Gd-chelate injection can be used to distinguish non-enhancing post-therapy hygroma, seroma, or hematoma from enhancing tumor recurrence, post-therapy fibrosis, granulation tissue, or inflammatory masses (Fig. 26.5). The T1 fat-suppressed images obtained after administration of the contrast agent should be compared to subtraction images (for instance, obtained in the dynamic sequence, see below) or to pre-contrast T1 fat-suppressed images to correct for pseudo-enhancement caused by scaling effects. Absence of contrast enhancement excludes recurrent tumor. On these standard contrast-enhanced images, the differentiation between recurrent viable tumor and post-therapy fibrosis or inflammatory pseudomasses may remain difficult. However, in these cases dynamic contrast-enhanced MR imaging may prove helpful [34, 51]. Dynamic contrast-enhanced MR imaging allows differentiation between inflammation and recurrent or residual tumor. After a rapid bolus injection of contrast, viable tumor exhibits rapid progressive increase of signal intensity followed by washout or plateau phase, whereas the signal from inflammatory changes will also increase but later [52] (Fig. 26.6). The exception is young granulation issue and reactive changes secondary to surgery or radiation within the last 6 months.

Diffusion-weighted imaging has potential in adding specificity in the diagnosis of recurrent tumor. Preliminary studies have shown a statistical difference between recurrent tumor and nonneoplastic masses secondary to earlier treatment [34, 51].

Each case of a suspicious (recurrent) mass should be treated as if it is a new sarcoma. Confirmation should be obtained by cytological sampling or core-needle (Tru-Cut or Jamshidi) biopsy after locoregional restaging by MR imaging. Histological biopsy should always be performed after MR imaging because reactive changes, hemorrhage, and edema secondary to biopsy may hamper interpretation of MR images and therefore interfere with staging.

Preliminary studies of positron emission tomography (PET) demonstrate the potential of PET with FDG as an additional tool for detecting local recurrence of soft tissue sarcoma. Moreover, PET seems particularly useful in patients with extensive histories of surgery and radiation therapy, in the setting in which MR imaging interpretation can be difficult [53, 54]. The caveat here is that a negative PET-CT can only be used to exclude recurrence, if the initial tumor accumulates sufficient FDG.

Desmoid-type fibromatoses are benign fibroblastic proliferations that arise in the deep soft tissues and are characterized by infiltrative growth in the surrounding soft tissue structures and the absence of a pseudocapsule. Because of this growth pattern, local recurrence after surgery is virtually unavoidable. Routine follow-up MR imaging of patients with desmoid-type fibromatosis seems justified not only to detect (often asymptomatic) local recurrence but also to evaluate the natural behavior of these lesions as they change from cellular active lesions toward mature, collagenous inert lesions [55] (Fig. 26.7).