Abstract
Objective:
We sought to define the prognostic significance of histologic subtype for extremity/truncal liposarcoma (LPS).
Summary of Background Data:
LPS, the most common sarcoma, is comprised of five histologic subtypes. Despite their distinct behaviors, LPS outcomes are frequently reported as a single entity.
Methods:
We analyzed data on all patients from a single-institution prospective database treated from July 1982 to September 2017 for primary, non-metastatic, extremity or truncal LPS of known subtype. Clinicopathologic variables were tested using competing risk analyses for association with disease-specific death (DSD), distant recurrence (DR), and local recurrence (LR).
Results:
Among 1001 patients, median follow-up in survivors was 5.4 years. Tumor size and subtype were independently associated with DSD and DR. Size, subtype, and R1 resection were independently associated with LR. DR was most frequent among pleomorphic and round cell LPS; the former recurred early (43% by 3 years), and the latter over a longer period (23%, 3 years; 37%, 10 years). LR was most common in dedifferentiated LPS, in which it occurred early (24%, 3 years; 33%, 5 years), followed by pleomorphic LPS (18%, 3 years; 25%, 10 years).
Conclusions:
Histologic subtype is the factor most strongly associated with DSD, DR, and LR in extremity/truncal LPS. Both risk and timing of adverse outcomes vary by subtype. These data may guide selective use of systemic therapy for patients with round cell and pleomorphic LPS, which carry a high risk of DR, and radiotherapy for LPS subtypes at high risk of LR when treated with surgery alone.
Mini-Abstract
In an analysis of 1001 patients with extremity/truncal liposarcoma, we characterize the distinct clinical behaviors of the five histologic subtypes of liposarcoma. Histologic subtype is the dominant prognostic factor for these tumors and dictates risk, pattern, and timing of recurrence and death from disease.
Introduction
Liposarcoma (LPS) is the most common type of soft tissue sarcoma, comprising approximately 20% of sarcomas in adults.1 Reported recurrence and survival rates vary widely among studies of LPS, largely due to variations in anatomic location and histologic subtype within the study populations.2–8 In the retroperitoneum the most common histologies are well- and dedifferentiated LPS, and local recurrence is the major cause of disease-specific mortality.1, 9 In contrast, in the extremity and trunk the full spectrum of LPS histologic subtypes are observed, and while local recurrence may precede distant disease, distant recurrence is the primary cause of disease-specific mortality.6, 8, 10
The World Health Organization (WHO) classifies LPS into four histologic subtypes: atypical lipomatous tumor (ALT, the term for well-differentiated LPS in the extremity and trunk), dedifferentiated, myxoid, and pleomorphic.11 ALT/well-differentiated LPS and dedifferentiated LPS are characterized by 12q amplification.12 Myxoid LPSs frequently contain a balanced translocation, t(12;16)(q13;p11), that fuses FUS to DDIT313, 14 and can be further subclassified into myxoid LPS (low grade, < 5% round cells) and round cell LPS (high grade, ≥ 5% round cells).15, 16 Pleomorphic LPSs typically have multiple regions of significant copy number amplification and deletion and complex chromosomal rearrangements.17 Each subtype is characterized by distinct morphology, molecular markers, and clinical course.12–14, 17.
In the extremity, ALT is the most common subtype of LPS.2, 18 Its high-grade counterpart, dedifferentiated LPS, is infrequently seen in the extremity, but is associated with higher rates of both local and distant recurrence compared with ALT.19 Myxoid LPS and the high-grade round cell LPS represent the second most common LPS subtype; the percentage of round cells is strongly associated with prognosis.6, 15, 20 These tumors are especially likely to metastasize to unusual sites, such as bone and fat pads in the abdomen and axilla.21–23 Pleomorphic LPS is the rarest LPS subtype and is defined by the presence of pleomorphic lipoblasts.24 This aggressive form of LPS most frequently arises in the extremities of older individuals and has a high propensity for metastasis.25, 26
Importantly, the national guidelines consider only grade and not histologic subtype in treatment recommendations.27 Histologic type and subtype are increasingly recognized as critical determinants of prognosis for soft tissue sarcoma, but careful study of individual subtypes has been limited by the rarity of the disease and need for long-term follow-up. For example, our institution has previously published two nomograms that include extremity and truncal LPS. The first did not stratify patients by LPS subtype.2 In the second, long-term follow-up was relatively limited, particularly for the less common LPS subtypes; and the inclusion of retroperitoneal LPS, which has distinct recurrence patterns from those of extremity and truncal LPS, complicated the outcome analysis.3 We therefore sought to define the prognostic factors, patterns, and timing of recurrence in a large cohort of extremity and truncal LPS. We hypothesized that, as in other similarly refined studies in sarcoma, histologic subtype would be a dominant factor influencing the outcomes of patients with LPS.5, 28
Methods
Patient Selection
The prospectively maintained sarcoma database at Memorial Sloan Kettering Cancer Center (MSKCC) was queried for patients with extremity and truncal LPS treated from July 1, 1982 until October 1, 2017. Only patients who underwent definitive surgical resection at MSKCC for primary LPS were included. After re-review by a sarcoma pathologist (M.H.), patients with unknown or undefinable histologic subtype were excluded (n=10). This study was undertaken after obtaining approval from the MSKCC Institutional Review Board.
Clinicopathologic Variables
Time to disease-specific death (DSD), first distant recurrence (DR), and first local recurrence (LR) were measured from the date of definitive surgery until the time of the event or last follow-up. Patients were followed clinically and radiographically as determined by the treating physicians. The typical follow-up protocol for low-grade sarcomas includes examination and scans at 6-month intervals for the first 5 years, yearly for years 5–10, and every 2 years thereafter. High-grade sarcomas are followed more intensively at 4-month intervals for the first 3 years, but subsequently in a similar fashion to low-grade tumors.
LPS was classified into 5 histologic subtypes based upon the WHO classification: atypical lipomatous tumors (ALT), dedifferentiated LPS, myxoid LPS (< 5% round cell component), round cell LPS (≥ 5% round cell), or pleomorphic LPS.11 Additional clinicopathologic variables included age (continuous and dichotomized at the median), sex, tumor site (upper extremity, lower extremity, trunk), tumor size (continuous and categorized as ≤ 5 cm, > 5-≤ 10 cm, or > 10 cm), depth, bone invasion (defined by pathologic review), nerve invasion (defined by pathologic review), resection type (limb-sparing or amputation), resection margin, and receipt of perioperative chemotherapy or radiation. Although the use of chemotherapy and/or radiation are described, these variables were not included in statistical analyses given the profound selection bias associated with the decision to administer additional therapy.
Statistical Analyses
Statistical associations with histologic subtype were examined using Fisher’s exact test for categorical variables and Kruskal-Wallis for continuous variables. Cumulative incidences were assessed for DSD, LR, and DR. Cumulative incidences were estimated, tested, and modeled in a competing risks framework using Gray’s k-test and the Fine and Gray regression model.29, 30 For analysis of DSD, any deaths unrelated to sarcoma were treated as competing events. Competing events for LR were deaths without LR, and competing events for DR were deaths without DR. For the analysis of LR, patients with R2 margins (n=21) and those who received radiation (n=313) were excluded (total of 323 patients excluded). Although the group undergoing radiation was highly selected, they were excluded to yield risk estimates that could inform the decision for adjuvant radiation.
Nomograms were developed for DSD and LR to estimate risk at 3, 5, and 10 years post-surgery. Nomograms were evaluated by calibration plots (including all patients or only those who received radiation) as well as concordance index, as previously described.5 All analyses were performed using SAS 9.4 and R 3.4.3 software. A p-value < 0.05 was considered statistically significant.
Results
Patient Characteristics
A total of 1001 patients with extremity or truncal LPS were included. The most common histologic subtype was atypical lipomatous tumor (ALT, n = 452), followed by myxoid LPS (n = 239), round cell LPS (n = 126), pleomorphic LPS (n = 111), and dedifferentiated LPS (n = 73). Myxoid and round cell tumors presented at younger median ages (40 and 46 years, respectively) than the other histologic subtypes (p < 0.001) (Table 1). ALTs had the largest median size (15 cm). The relative proportions of the 5 histologic subtypes varied among anatomic sites (p < 0.001), although the lower extremity was the most common site for all histologies (75%). Nerve (0.9%) and bone (1.0%) invasion were rare. Grossly positive margins (R2 resections) were uncommon across all histologic subtypes (2.1%). The frequency of microscopically positive margins varied among histologic subtypes (p < 0.001) and were most frequently seen in ALT (20%) and dedifferentiated LPS (29%).
Table 1.
Clinicopathologic characteristics of patients with extremity and truncal liposarcoma. Values are n (%) unless otherwise stated.
| Total (n=1,001) | Atypical lipomatous tumor (n=452) | Myxoid (n=239) | Round cell (n=126) | Pleomorphic (n=111) | Dedifferentiated (n=73) | p value | ||
|---|---|---|---|---|---|---|---|---|
| Age | Median (range) | 55 (13–95) | 59 (18–93) | 40 (15–82) | 46 (23–88) | 57 (13–95) | 66 (24–95) | <.001 |
| Age | ≤ 54 | 500 (50) | 170 (37.6) | 184 (77) | 92 (73) | 38 (34.2) | 16 (21.9) | <.001 |
| > 54 | 501 (50) | 282 (62.4) | 55 (23) | 34 (27) | 73 (65.8) | 57 (78.1) | ||
| Sex | Female | 424 (42.4) | 197 (43.6) | 107 (44.8) | 45 (35.7) | 47 (42.3) | 28 (38.4) | 0.465 |
| Male | 577 (57.6) | 255 (56.4) | 132 (55.2) | 81 (64.3) | 64 (57.7) | 45 (61.6) | ||
| Anatomic site | Lower extremity | 747 (74.6) | 316 (69.9) | 199 (83.3) | 112 (88.9) | 70 (63.1) | 50 (68.5) | <.001 |
| Trunk | 128 (12.8) | 60 (13.3) | 23 (9.6) | 9 (7.1) | 21 (18.9) | 15 (20.5) | ||
| Upper extremity | 126 (12.6) | 76 (16.8) | 17 (7.1) | 5 (4) | 20 (18) | 8 (11) | ||
| Tumor size | Median (range) | 12.4 (1.00–65.0) | 15.0 (1.00–58.0) | 10.0 (1.50–33.0) | 11.6 (3.50–36.0) | 10.0 (1.00–65.0) | 12.0 (3.10–35.0) | <.001 |
| Tumor size | ≤ 5 cm | 132 (13.3) | 48 (10.7) | 41 (17.3) | 14 (11.3) | 21 (18.9) | 8 (11) | <.001 |
| > 5, ≤10 cm | 268 (26.9) | 89 (19.8) | 83 (35) | 38 (30.6) | 43 (38.7) | 15 (20.5) | ||
| > 10 cm | 595 (59.8) | 313 (69.6) | 113 (47.7) | 72 (58.1) | 47 (42.3) | 50 (68.5) | ||
| Unknown | 6 (.) | 2 (.) | 2 (.) | 2 (.) | 0 (.) | 0 (.) | ||
| Depth | Superficial | 119 (11.9) | 44 (9.7) | 35 (14.6) | 10 (7.9) | 18 (16.2) | 12 (16.4) | 0.054 |
| Deep | 882 (88.1) | 408 (90.3) | 204 (85.4) | 116 (92.1) | 93 (83.8) | 61 (83.6) | ||
| Limb-sparing surgery | Yes | 990 (99) | 451 (99.8) | 237 (99.2) | 121 (96) | 109 (99.1) | 72 (98.6) | 0.009 |
| No | 10 (1) | 1 (0.2) | 2 (0.8) | 5 (4) | 1 (0.9) | 1 (1.4) | ||
| Unknown | 1 (.) | 0 (.) | 0 (.) | 0 (.) | 1 (.) | 0 (.) | ||
| Margin | R0 | 818 (81.8) | 349 (77.4) | 218 (91.2) | 107 (84.9) | 92 (82.9) | 52 (71.2) | <.001 |
| R1 | 161 (16.1) | 91 (20.2) | 19 (7.9) | 16 (12.7) | 14 (12.6) | 21 (28.8) | ||
| R2 | 21 (2.1) | 11 (2.4) | 2 (0.8) | 3 (2.4) | 5 (4.5) | 0 (0) | ||
| Unknown | 1 (.) | 1 (.) | 0 (.) | 0 (.) | 0 (.) | 0 (.) | ||
| Neural invasion | No | 992 (99.1) | 448 (99.1) | 237 (99.2) | 126 (100) | 110 (99.1) | 71 (97.3) | 0.383 |
| Yes | 9 (0.9) | 4 (0.9) | 2 (0.8) | 0 (0) | 1 (0.9) | 2 (2.7) | ||
| Bone invasion | No | 991 (99) | 452 (100) | 234 (97.9) | 123 (97.6) | 109 (98.2) | 73 (100) | 0.004 |
| Yes | 10 (1) | 0 (0) | 5 (2.1) | 3 (2.4) | 2 (1.8) | 0 (0) | ||
| Perioperative chemotherapy | None | 894 (89.3) | 445 (98.2) | 223 (93.3) | 89 (70.6) | 73 (65.8) | 64 (87.7) | <.001 |
| Adjuvant | 60 (6) | 6 (1.3) | 7 (2.9) | 17 (13.5) | 24 (21.6) | 6 (8.2) | ||
| Neoadjuvant | 47 (4.7) | 1 (0.2) | 9 (3.8) | 20 (15.9) | 14 (12.6) | 3 (4.1) | ||
| Perioperative radiation | None | 688 (68.7) | 413 (91.4) | 134 (56.1) | 52 (41.3) | 40 (36) | 41 (56.2) | <.001 |
| Adjuvant | 269 (26.9) | 37 (8.2) | 83 (34.7) | 64 (50.8) | 64 (57.7) | 29 (39.7) | ||
| Neoadjuvant | 44 (4.4) | 2 (0.4) | 22 (9.2) | 10 (7.9) | 7 (6.3) | 3 (4.1) | ||
| Follow-up (months) | Median (range) | 64.7 (0.03–398) | 57.3 (0.03–342) | 78.1 (0.07–398) | 95.5 (0.07–267) | 69.0 (0.49–379) | 62.3 (0.30–316) | |
The majority of patients were treated with surgery alone (65%). Amputation was rarely necessary at the time of presentation (1%). Radiation therapy use varied by histologic subtype (p < 0.001); it was commonly employed in the high-grade subtypes (pleomorphic, 64%; round cell, 59%) but not in ALT (9%). Similarly, rates of chemotherapy use varied (p < 0.001); chemotherapy was utilized in 34% of patients with pleomorphic LPS and 29% of patients with round cell LPS, but in only 1% of patients with ALT (all but one of the ALT patients receiving chemotherapy were treated in the adjuvant setting in combination with brachytherapy in the 1980s).
Disease-Specific Death
With a median follow-up in survivors of 65 months, 248 patients died during follow-up, 103 (42%) of whom died from LPS. For the entire cohort, the cumulative rate of DSD at 3 and 10 years was 7% (95% CI, 5–8%) and 13% (95% CI, 11–16%), respectively (Fig. 1A).
Figure 1. Cumulative incidence and timing of disease-specific death vary by histologic subtype.
A) Disease-specific death (DSD) vs. death from other causes for the entire cohort. B) Cumulative incidence of DSD according to histologic subtype. C) Timing of DSD, displayed as a histogram showing the percentage of events occurring by 3 years and 10 years.
Disease-specific death from extremity/truncal ALTs was extremely rare (1.5% at 10 years; 95% CI, 0.5–3.7%). In contrast, the cumulative incidence of DSD at 10 years for pleomorphic LPS was 37.7% (95% CI, 27.9–47.4%) and for round cell LPS was 28.3% (95% CI, 19.3–38.0%). DSD risk was intermediate for dedifferentiated LPS (12.6% DSD at 10 years; 95% CI, 4.7–24.4%) and myxoid LPS (12.2% DSD at 10 years; 95% CI, 7.6–18%) (Fig. 1B). Death from pleomorphic LPS occurred early, whereas over half of the deaths from myxoid and round cell LPS occurred more than 3 years after diagnosis (Fig. 1C).
The prognostic importance of tumor size varied by histologic subtype (Fig. 2). Across all histologic subtypes, tumors ≤ 5 cm were rarely associated with DSD. In pleomorphic LPS, and to a lesser degree in myxoid LPS, size cut-offs of ≤ 5 cm, > 5- ≤ 10 cm, and > 10 cm stratified patients’ risk of DSD. Among patients with round cell LPS, only a small difference in DSD was observed between tumors > 5- ≤ 10 cm and those > 10 cm (5-year DSD: 14%; 95% CI, 4–30%, and 22%; 95% CI, 12–32%, respectively). Patients with tumors > 10 cm, however, had a higher incidence of DSD in the first 3 years following diagnosis compared to those with 5- to 10-cm tumors. Patients with round cell LPS ≤ 5 cm had an excellent prognosis (5-year DSD: 0%; 95% CI, n/a). In patients with dedifferentiated LPS, DSD was not observed if tumors were < 10 cm. Patients with ALTs had good outcome regardless of tumor size.
Figure 2. Influence of size on disease-specific death varies by histologic subtype.
A) Pleomorphic LPS. B) Round cell LPS. C) Myxoid LPS. D) Dedifferentiated LPS. E) ALT.
Distant Recurrence
During follow-up 137 patients developed DR, and in that time 142 patients died from other causes. For the entire cohort, the cumulative rate of DR at 3 and 10 years was 12% (95% CI, 10–14%) and 17% (95% CI, 14–20%), respectively (Fig. 3A).
Figure 3. Cumulative incidence and timing of distant recurrence vary by histologic subtype.
A) Distant recurrence vs. death prior to recurrence for the entire cohort. B) Cumulative incidence of distant recurrence according to histologic subtype. C) Timing of distant recurrence, displayed as a histogram showing the percentage of events occurring by 3 years and 10 years.
The cumulative incidences of DR varied similarly among subtypes as those of DSD. The greatest incidence of DR was in pleomorphic LPS (44.6% at 10 years; 95% CI, 34.5–54.3%) and round cell LPS (37.4%; 95% CI, 27.7–47.0%), intermediate in dedifferentiated LPS (18.0% DR; 95% CI, 9.5–28.8%) and myxoid LPS (17.7% DR; 95% CI, 12.3–23.8%), and the lowest in ALT (1.9% at 10 years; 95% CI, 0.6–4.4%) (Fig. 3B). All but one DR for pleomorphic LPS occurred within the first 3 years. In contrast, DR for both round cell and myxoid LPS occurred steadily throughout the follow-up period, with approximately 40% of all DR occurring after 3 years (Fig. 3C).
Of the 137 patients who developed DR, isolated lung recurrence was most frequent (39%), followed by soft tissue (18%), other sites (15%), bone (14%), or multiple synchronous sites (13%). The anatomic site of distant recurrence varied by histologic subtype (p < 0.001). Most (76%) metastases from pleomorphic LPS were to the lung, compared with 46% of those from dedifferentiated, 20% of round cell, and 14% of myxoid LPS. Only 5 distant metastases were observed in patients with ALTs. Both myxoid and round cell tumors frequently metastasized to bone (17% and 27%, respectively), soft tissue (20% and 29%), and other sites (31% and 12%) (Supplemental Tables 1 and 2).
Local Recurrence
After excluding patients with an R2 resection, the cumulative incidence of LR in the overall cohort at 3 and 10 years was 8% (95% CI, 6–9%) and 13% (95% CI, 10–15%), respectively. LR incidence was similar in the 314 patients selected for radiation therapy: 6.9% (95% CI, 4.4–10.2%) at 3 years and 11.5% (95% CI, 8.0–15.7%) at 10 years. To minimize selection bias and to allow for estimation of risk of LR without treatment, patients undergoing radiation were excluded from subsequent analyses of LR.
Among the remaining 678 patients, 72 patients developed a LR during follow-up, and 89 patients died prior to LR. In this non-radiated cohort, the cumulative rate of LR at 3 and 10 years was 7.9% (95% CI, 5.8–10.4%) and 13.5% (95% CI, 10.4–17.0%), respectively (Fig. 4A).
Figure 4. Cumulative incidence and timing of local recurrence vary by histologic subtype.
A) Local recurrence vs. death prior to recurrence for non-radiated patients. B) Cumulative incidence of local recurrence according to histologic subtype. C) Timing of local recurrence, displayed as a histogram showing the percentage of events occurring by 3 years and 10 years. Analysis excludes patients undergoing R2 resection or receiving radiation.
The incidence of LR at 10 years was substantially higher for dedifferentiated LPS (32.6%; 95% CI, 16.3–50.1%) than for any other histologic subtype. LR rates for the remainder of the histologic subtypes ranged from 9.5% (ALT; 95% CI, 5.9–14.0%) to 24.8% (pleomorphic LPS; 95% CI, 11.3–41.0%) at 10 years (Fig. 4B). LR occurred early in dedifferentiated LPS and pleomorphic LPS (74% and 71% of recurrences within the first 3 years, respectively) (Fig. 4C). Early LR was less common for ALT and myxoid/round cell LPS, but delayed LR was more frequent. For ALT in particular, recurrences continued beyond 10 years such that rates approached those of pleomorphic LPS by 15 years (Fig. 4B).
Analysis of Factors Predictive of DSD, DR, and LR
On univariable analysis, the clinicopathologic factors significantly associated with risk of DSD and DR were histologic subtype, larger tumor size, and greater tumor depth. On multivariable analysis, histologic subtype and tumor size remained significantly associated with both outcomes (each p < 0.001), while tumor depth did not (Table 2). Univariable analysis identified histologic subtype, older patient age, larger tumor size, greater tumor depth, and R1 resection as significantly associated with LR. In a multivariable analysis, histologic subtype (p < 0.001), resection margin (p = 0.001), and tumor size (p = 0.003) remained independently associated with LR (Table 2).
Table 2.
Multivariable analyses of clinical and pathological variables associated with DSD, DR, and LR.
| Variable | Disease-specific death | Distant recurrence | Local recurrence* | |||
|---|---|---|---|---|---|---|
| Multivariable HR (95% CI) | p value | Multivariable HR (95% CI) | p value | Multivariable HR (95% CI) | p value | |
| Histologic subtype | <0.001 | <0.001 | 0.001 | |||
| Atypical lipomatous tumor (n=452) | Ref | -- | Ref | -- | Ref | -- |
| Dedifferentiated (n=73) | 6.6 (2.2–19.8) | <0.001 | 17.0 (5.8–49.4) | <0.001 | 5.1 (2.4–11.0) | <0.001 |
| Myxoid (n=239) | 8.7 (3.7–20.3) | <0.001 | 17.9 (6.9–46.3) | <0.001 | 2.4 (1.2–4.9) | 0.013 |
| Round cell (n=127) | 17.1 (7.5–39.1) | <0.001 | 38.7 (15.2–98.9) | <0.001 | 1.6 (0.6–4.2) | 0.341 |
| Pleomorphic (n=111) | 34.3 (14.8–79.9) | <0.001 | 65.0 (25.0–168.7) | <0.001 | 4.5 (1.8–11.1) | 0.001 |
| Tumor size** | 1.9 (1.5–2.4) | <0.001 | 1.8 (1.5–2.2) | <0.001 | 1.6 (1.2–2.1) | <0.001 |
| Resection margin | ||||||
| R0 (n=818) | Ref | -- | Ref | -- | Ref | -- |
| R1 (n=161) | NS*** | NS*** | NS*** | NS*** | 2.3 (1.3–3.9) | 0.002 |
Patients with R2 resection and perioperative radiation (n=335) were excluded from LR analysis.
Tumor size as a continuous variable, log2 transformed for normalization.
Not significant on univariable analysis, thus not included in multivariable model.
Nomograms incorporating tumor size as a continuous variable were developed for the prediction of DSD and LR at 3, 5, and 10 years for patients treated with surgery alone (Supplemental Figs. 1A and 1B, respectively). The concordance indices at 3, 5, and 10 years were 0.82, 0.84, and 0.81 for the DSD model, and 0.77, 0.74, and 0.71 for the LR model. Calibration plots demonstrate good agreement between nomogram-predicted and observed events (Supplemental Figs. 2–3). To determine whether the LR nomogram is relevant for predicting the outcomes of patients preoperatively treated with radiation, we also generated a calibration plot using data from these patients (Supplemental Fig. 4). The nomogram was less accurate in predicting LR in this group (concordance indices at 3, 5, and 10 years: 0.61, 0.63, 0.63, respectively), consistent with the competing effects of patient selection and the efficacy of radiation therapy, but still performed fairly well, particularly at lower levels of risk.
Discussion
Based on long-term data from a large population of patients with primary extremity and truncal LPS treated with surgery at a single institution, we identify histologic subtype as the dominant driver of the risk and timing of DSD and recurrence. In fact, in this site-restricted cohort, tumor size was the only additional factor associated with risk of DSD.
In a previous analysis of patients treated for LPS at our institution, dedifferentiated and, to a lesser degree, ALT/well-differentiated LPS were associated with substantially worse survival than observed in this study.3 This difference results primarily from the prior study’s inclusion of patients with retroperitoneal tumors, in whom local recurrence often directly causes mortality.9 As such, the anatomic location of an LPS is a critically important determinant of treatment approach and outcome.3, 9, 31, 32 In the extremity, where local recurrence may be highly morbid but is far less likely to be fatal, the 10-year DR rate (18%) and DSD rate (12.6%) are low for dedifferentiated LPS.
We observe distinct oncologic outcomes for specific histologic subtypes, even of the same grade classification. For example, the rates of DR and DSD in dedifferentiated LPS, a high-grade tumor, are nearly identical to those observed in myxoid LPS, a low-grade tumor. Thus, the kinetics of recurrence in dedifferentiated LPS are similar to those of a high-grade tumor (such as pleomorphic LPS), but the absolute rate of distant events is more consistent with the low-grade subtypes. The low rate of distant events may also reflect the influence of tumor size in dedifferentiated LPS. In the extremity we only observed DSD in patients with tumors > 10 cm. This may represent an important threshold in dedifferentiated LPS; patients with this subtype of tumors that are larger than 10 cm may benefit most from novel targeted or immune therapeutics provided in (neo)adjuvant trials. The relatively low risk of DR associated with dedifferentiated LPS (18% cumulative incidence at 10 years), almost identical to that of myxoid liposarcoma, a low-grade LPS, represents another unexpected finding from this study. Although pleomorphic and round cell LPS, both considered high-grade subtypes, had the same cumulative incidence of DR (45% at 15 years), virtually all patients with pleomorphic LPS experienced DR within 3 years of diagnosis, compared with about 50% of those with round cell LPS, whose DRs occurred 3 to 15 years after initial diagnosis. These findings highlight the importance of LPS histologic subtype in more precisely defining both the risk and kinetics of DR compared to grade alone, and emphasize the need for continuing to collect and analyze increasingly large and homogenous subpopulations of patients with sarcoma.
Despite the low cumulative incidence of DR in dedifferentiated LPS, the risk of LR is substantial, consistent with its high grade. Among patients treated with surgery alone, 33% (95% CI, 16.3–50.1%) developed local recurrence, all within 5 years. Interestingly, in the patients who did receive radiation, the 5-year LR rate was lower: 22% (95% CI, 8.4–38.7%). The slightly improved outcome observed in the radiated group, who were selected for radiation based upon the prediction of increased risk of recurrence, may support the use of radiation in dedifferentiated LPS. Given that 67% of patients did not have a local recurrence, however, one must balance the potential benefit with known side effects. The kinetics of LR in dedifferentiated LPS are distinct from its low-grade counterpart, ALT, in which 75% of local recurrences for ALT occur after 5 years. This extended risk of LR makes long-term follow-up necessary, including beyond the 10-year time point recommended by the NCCN.27
In contrast to ALT and dedifferentiated LPS, round cell LPS and its low-grade counterpart, myxoid LPS, demonstrate similar kinetics for DSD, DR, and LR. However, the absolute risk of distant events is significantly greater in round cell LPS. For both myxoid and round cell subtypes, approximately 60% of DSD and 40% of DR occur beyond 3 years, making long-term follow-up with yearly chest, abdomen, and pelvic CT scans potentially valuable for these patients.
The influence of tumor size on outcome also appears to be subtype-specific. Tumor size dramatically stratifies the risk of patients with pleomorphic LPS for DSD (6% if ≤ 5 cm, 16% if > 5–10 cm, and 46% if > 10 cm at 3 years), whereas in round cell LPS, 5 cm appears to be the only important size threshold. As mentioned above, in dedifferentiated LPS, death from disease is only observed in patients with tumors larger than 10 cm. This finding is particularly noteworthy given the recent changes to the staging system in the 8th edition of the AJCC to create four different T stages (≤ 5 cm, > 5–10 cm, > 10–15 cm, > 15 cm) for extremity sarcomas.33 The varying influence of tumor size by histologic subtype highlights the challenge of applying a single staging system across sarcoma and reinforces the need for histology and site-specific nomograms.3, 5, 34
Our findings also have implications for the use of (neo)adjuvant chemotherapy in LPS. Although the ability of chemotherapy to reduce DSD and DR in LPS remains unproven and a subject of debate, the data presented here clearly demonstrate that patients with round cell and pleomorphic LPSs > 5 cm have the greatest potential to benefit. Patients with dedifferentiated LPS, however, have a relatively low incidence of DSD (18% at 10 years). That finding, as well as the known low response rate to chemotherapy, makes patients with dedifferentiated LPS poor candidates for (neo)adjuvant therapy.35, 36
These findings highlight the importance of considering specific histologic subtypes, as opposed to grade and size alone, in LPS management. Prior studies have broadly characterized the behaviors associated with each histologic subtype described here.6, 18–22, 24–26, 35 Yet smaller size and shorter follow-up have limited the reliability of those descriptions. With this large experience, we hope to provide accurate estimates that can be used to confidently counsel patients, guide evidence-based follow-up regimens, and present baseline outcomes that can be used in the development of future prospective studies.
There are limitations to this study. The data was collected prospectively for our sarcoma database, but this is a retrospective analysis and carries the associated limitations. The single-center nature of the data may limit the findings’ generalizability, but this concern must be weighed against the benefits of a single-center experience, including consistent pathologic evaluation and relatively homogeneous management. The long period of data collection included in this study may also raise concern that these findings reflect historic outcomes that are no longer relevant. To account for possible improvements in outcome due to changes in treatment, we analyzed DSS by decade and found that it did not improve over time for any histologic subtype, nor was decade of treatment an independent predictor of this outcome (data not shown).
Perhaps the most significant potential confounder is the selection bias regarding administration of multimodality therapy. Our institution selectively employs both chemotherapy and radiation for sarcoma treatment, for which the criteria have evolved over time as our understanding of these tumors has improved. Overall neoadjuvant or adjuvant chemotherapy were rarely given to patients with LPS (11%) but were much more frequently used for the high-grade subtypes. However, numerous clinical trials have shown that the impact of adjuvant chemotherapy is, at best, small.37, 38 Thus, the impact of chemotherapy on the DSD and DR estimates of the cohort is unlikely to be substantial.
Radiation therapy is likely to have more significantly affected the LR estimates described here. Radiation has been associated with a reduction in local recurrence of approximately 50% and was utilized in a substantial proportion of patients (31%), particularly in the high-grade subtypes.39, 40 For this reason, we chose to exclude patients who received radiation treatment from the LR multivariable and nomogram analysis. The exclusion allows us to provide a natural history of the risk associated with these lesions that can be used to inform the decision for adjuvant radiation, but unfortunately excludes the patients likely to be at highest risk of recurrence.
Conclusion
Histologic subtype is the factor most significantly associated with the risk and timing of recurrence and disease-specific death in truncal/extremity LPS. The distinct behaviors of the various subtypes may inform the selective use of (neo)adjuvant chemo- and radiotherapy.
Supplementary Material
Acknowledgment
The authors would like to thank Jessica Moore for her editorial assistance with the manuscript and figures. This work was supported in part by the NIH SPORE in Soft Tissue Sarcoma P50 CA217694, NIH/NCI R01 CA158301, the Liposarcoma Fund, the Kristen Ann Carr Fund, and the NIH/NCI Cancer Center Support Grant P30 CA008748.
Sources of support: This work was financially supported by the NIH SPORE in Soft Tissue Sarcoma P50 CA140146, the Liposarcoma Fund, the Kristen Ann Carr Fund, and the NIH/NCI Cancer Center Support Grant P30 CA008748.
References
- 1.Brennan MF, Antonescu CR, Moraco N, et al. Lessons learned from the study of 10,000 patients with soft tissue sarcoma. Ann Surg 2014; 260(3):416–21; discussion 421–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cahlon O, Brennan MF, Jia X, et al. A postoperative nomogram for local recurrence risk in extremity soft tissue sarcomas after limb-sparing surgery without adjuvant radiation. Ann Surg 2012; 255(2):343–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Dalal KM, Kattan MW, Antonescu CR, et al. Subtype specific prognostic nomogram for patients with primary liposarcoma of the retroperitoneum, extremity, or trunk. Ann Surg 2006; 244(3):381–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Linehan DC, Lewis JJ, Leung D, et al. Influence of biologic factors and anatomic site in completely resected liposarcoma. J Clin Oncol 2000; 18(8):1637–43. [DOI] [PubMed] [Google Scholar]
- 5.Tan MC, Brennan MF, Kuk D, et al. Histology-based Classification Predicts Pattern of Recurrence and Improves Risk Stratification in Primary Retroperitoneal Sarcoma. Ann Surg 2016; 263(3):593–600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gronchi A, Miceli R, Fiore M, et al. Extremity soft tissue sarcoma: adding to the prognostic meaning of local failure. Ann Surg Oncol 2007; 14(5):1583–90. [DOI] [PubMed] [Google Scholar]
- 7.Gronchi A, Miceli R, Allard MA, et al. Personalizing the approach to retroperitoneal soft tissue sarcoma: histology-specific patterns of failure and postrelapse outcome after primary extended resection. Ann Surg Oncol 2015; 22(5):1447–54. [DOI] [PubMed] [Google Scholar]
- 8.Pearlstone DB, Pisters PW, Bold RJ, et al. Patterns of recurrence in extremity liposarcoma: implications for staging and follow-up. Cancer 1999; 85(1):85–92. [DOI] [PubMed] [Google Scholar]
- 9.Canter RJ, Qin LX, Ferrone CR, et al. Why do patients with low-grade soft tissue sarcoma die? Ann Surg Oncol 2008; 15(12):3550–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eilber FC, Rosen G, Nelson SD, et al. High-grade extremity soft tissue sarcomas: factors predictive of local recurrence and its effect on morbidity and mortality. Ann Surg 2003; 237(2):218–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Fletcher CDM, World Health Organization., International Agency for Research on Cancer. WHO classification of tumours of soft tissue and bone. 4th ed. Lyon: IARC Press, 2013. [Google Scholar]
- 12.Barretina J, Taylor BS, Banerji S, et al. Subtype-specific genomic alterations define new targets for soft-tissue sarcoma therapy. Nat Genet 2010; 42(8):715–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Barone MV, Crozat A, Tabaee A, et al. CHOP (GADD153) and its oncogenic variant, TLS-CHOP, have opposing effects on the induction of G1/S arrest. Genes Dev 1994; 8(4):453–64. [DOI] [PubMed] [Google Scholar]
- 14.Kuroda M, Ishida T, Takanashi M, et al. Oncogenic transformation and inhibition of adipocytic conversion of preadipocytes by TLS/FUS-CHOP type II chimeric protein. Am J Pathol 1997; 151(3):735–44. [PMC free article] [PubMed] [Google Scholar]
- 15.Antonescu CR, Tschernyavsky SJ, Decuseara R, et al. Prognostic impact of P53 status, TLS-CHOP fusion transcript structure, and histological grade in myxoid liposarcoma: a molecular and clinicopathologic study of 82 cases. Clin Cancer Res 2001; 7(12):3977–87. [PubMed] [Google Scholar]
- 16.Moreau LC, Turcotte R, Ferguson P, et al. Myxoid\round cell liposarcoma (MRCLS) revisited: an analysis of 418 primarily managed cases. Ann Surg Oncol 2012; 19(4):1081–8. [DOI] [PubMed] [Google Scholar]
- 17.Taylor BS, Barretina J, Socci ND, et al. Functional copy-number alterations in cancer. PLoS One 2008; 3(9):e3179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kooby DA, Antonescu CR, Brennan MF, et al. Atypical lipomatous tumor/well-differentiated liposarcoma of the extremity and trunk wall: importance of histological subtype with treatment recommendations. Ann Surg Oncol 2004; 11(1):78–84. [DOI] [PubMed] [Google Scholar]
- 19.Henricks WH, Chu YC, Goldblum JR, et al. Dedifferentiated liposarcoma: a clinicopathological analysis of 155 cases with a proposal for an expanded definition of dedifferentiation. Am J Surg Pathol 1997; 21(3):271–81. [DOI] [PubMed] [Google Scholar]
- 20.Hoffman A, Ghadimi MP, Demicco EG, et al. Localized and metastatic myxoid/round cell liposarcoma: clinical and molecular observations. Cancer 2013; 119(10):1868–77. [DOI] [PubMed] [Google Scholar]
- 21.Estourgie SH, Nielsen GP, Ott MJ. Metastatic patterns of extremity myxoid liposarcoma and their outcome. J Surg Oncol 2002; 80(2):89–93. [DOI] [PubMed] [Google Scholar]
- 22.Kilpatrick SE, Doyon J, Choong PF, et al. The clinicopathologic spectrum of myxoid and round cell liposarcoma. A study of 95 cases. Cancer 1996; 77(8):1450–8. [DOI] [PubMed] [Google Scholar]
- 23.Schwab JH, Boland P, Guo T, et al. Skeletal metastases in myxoid liposarcoma: an unusual pattern of distant spread. Ann Surg Oncol 2007; 14(4):1507–14. [DOI] [PubMed] [Google Scholar]
- 24.Downes KA, Goldblum JR, Montgomery EA, et al. Pleomorphic liposarcoma: a clinicopathologic analysis of 19 cases. Mod Pathol 2001; 14(3):179–84. [DOI] [PubMed] [Google Scholar]
- 25.Gebhard S, Coindre JM, Michels JJ, et al. Pleomorphic liposarcoma: clinicopathologic, immunohistochemical, and follow-up analysis of 63 cases: a study from the French Federation of Cancer Centers Sarcoma Group. Am J Surg Pathol 2002; 26(5):601–16. [DOI] [PubMed] [Google Scholar]
- 26.Hornick JL, Bosenberg MW, Mentzel T, et al. Pleomorphic liposarcoma: clinicopathologic analysis of 57 cases. Am J Surg Pathol 2004; 28(10):1257–67. [DOI] [PubMed] [Google Scholar]
- 27.National Comprehensive Cancer Network Guidelines for Soft Tissue Sarcoma 2019. Available at: https://www.nccn.org/. Accessed 4/30/2019.
- 28.Gronchi A, Strauss DC, Miceli R, et al. Variability in Patterns of Recurrence After Resection of Primary Retroperitoneal Sarcoma (RPS): A Report on 1007 Patients From the Multi-institutional Collaborative RPS Working Group. Ann Surg 2016; 263(5):1002–9. [DOI] [PubMed] [Google Scholar]
- 29.Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999; 94(446):496–509. [Google Scholar]
- 30.Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 1988:1141–1154. [Google Scholar]
- 31.Mussi C, Collini P, Miceli R, et al. The prognostic impact of dedifferentiation in retroperitoneal liposarcoma: a series of surgically treated patients at a single institution. Cancer 2008; 113(7):1657–65. [DOI] [PubMed] [Google Scholar]
- 32.Nijhuis PH, Sars PR, Plaat BE, et al. Clinico-pathological data and prognostic factors in completely resected AJCC stage I-III liposarcomas. Ann Surg Oncol 2000; 7(7):535–43. [DOI] [PubMed] [Google Scholar]
- 33.Amin MB, Edge S, Greene F, et al. AJCC cancer staging manual. 8th ed. Chicago, IL: Springer, 2017. [Google Scholar]
- 34.Raut CP, Miceli R, Strauss DC, et al. External validation of a multi-institutional retroperitoneal sarcoma nomogram. Cancer 2016; 122(9):1417–24. [DOI] [PubMed] [Google Scholar]
- 35.Italiano A, Toulmonde M, Cioffi A, et al. Advanced well-differentiated/dedifferentiated liposarcomas: role of chemotherapy and survival. Ann Oncol 2012; 23(6):1601–7. [DOI] [PubMed] [Google Scholar]
- 36.Jones RL, Fisher C, Al-Muderis O, et al. Differential sensitivity of liposarcoma subtypes to chemotherapy. Eur J Cancer 2005; 41(18):2853–60. [DOI] [PubMed] [Google Scholar]
- 37.Pervaiz N, Colterjohn N, Farrokhyar F, et al. A systematic meta-analysis of randomized controlled trials of adjuvant chemotherapy for localized resectable soft-tissue sarcoma. Cancer 2008; 113(3):573–81. [DOI] [PubMed] [Google Scholar]
- 38.Woll PJ, Reichardt P, Le Cesne A, et al. Adjuvant chemotherapy with doxorubicin, ifosfamide, and lenograstim for resected soft-tissue sarcoma (EORTC 62931): a multicentre randomised controlled trial. Lancet Oncol 2012; 13(10):1045–54. [DOI] [PubMed] [Google Scholar]
- 39.Pisters PW, Harrison LB, Leung DH, et al. Long-term results of a prospective randomized trial of adjuvant brachytherapy in soft tissue sarcoma. J Clin Oncol 1996; 14(3):859–68. [DOI] [PubMed] [Google Scholar]
- 40.Yang JC, Chang AE, Baker AR, et al. Randomized prospective study of the benefit of adjuvant radiation therapy in the treatment of soft tissue sarcomas of the extremity. J Clin Oncol 1998; 16(1):197–203. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.