Professional Documents
Culture Documents
Marco Fiore, MD1 BACKGROUND. The objective of this study was to investigate prognostic factors
Federica Grosso, MD2 and clinical outcome of myxoid/round cell and pleomorphic liposarcoma.
Salvatore Lo Vullo, Bsc3 METHODS. Three hundred twenty-nine patients with localized myxoid/round cell
Elisabetta Pennacchioli, MD1 or pleomorphic liposarcoma who underwent surgery at the Istituto Nazionale per
Silvia Stacchiotti, MD2 lo Studio e la Cura dei Tumori (Milan, Italy) over 25 years were reviewed. The
Andrea Ferrari, MD2 reates of local recurrence, distant metastases, and survival were studied.
Paola Collini, MD4 RESULTS. Two hundred fourteen patients presented with primary disease, and
Laura Lozza, MD5 115 patients had locally recurrent tumors. The disease-specific survival rate was
Luigi Mariani, MD3 75% at 10 years, and the local recurrence and distant metastases incidence were
Paolo G. Casali, MD2 25% and 15%, respectively. Presentation with recurrent disease, tumor size (>10
Alessandro Gronchi, MD1 cm), tumor grade (French Federation of Cancer Centers grade II or III vs grade I),
and positive surgical margins were independent predictors of death. Tumor site
1
Department of Surgery, National Cancer Insti- and radiation therapy also played a role, mostly related to their effect on local
tute, Milan, Italy. outcome. Pathologic grade and histologic subtype influenced distant metastases.
2
Department of Cancer Medicine, National Can- Extrapulmonary metastases were associated with poorer postmetastatic disease-
cer Institute, Milan, Italy. specific survival.
3 CONCLUSIONS. Myxoid/round cell liposarcomas shared similar prognostic factors
Department of Biostatistics, National Cancer
Institute, Milan, Italy. with other soft tissue sarcomas and had a relatively good clinical outcome. The
presence of >5% of round cell component singled out a group of patients at
4
Department of Pathology, National Cancer Insti-
greater risk of metastases and death but with a broad spectrum of disease aggres-
tute, Milan, Italy.
siveness. Extrapulmonary metastases were a peculiar pattern of myxoid/round
5
Department of Diagnostic Imaging and Radio- cell liposarcoma that require special consideration for treatment and prognosis.
therapy, National Cancer Institute, Milan, Italy.
Cancer 2007;109:2522–31. 2007 American Cancer Society.
tinguishable from RCLPS has been observed.5 operating surgeon. The margins were inked and were
Recently, a rare epithelioid variant of PLPS carrying sampled separately. The closest margin was categor-
the typical MLPS/RCLPS FUS-CHOP transcript also ized microscopically as positive (tumor within 1 mm
has been described.6 from the inked surface) or negative (absence of tumor
MLPS/RCLPS tumors are known as particularly within 1 mm from the inked surface).
sensitive to radiotherapy and chemotherapy compared Clinical data were extracted from a prospective
with other histologic subtypes of soft tissue sar- database of all adult patients with soft tissue sar-
coma.7,8 Recent data have raised new interest in coma who were treated at our institution. The data
medical treatments for MLPS/RCLPS.9 Therefore, we retrieved included sex, age at diagnosis, tumor site,
undertook the current retrospective analysis of our phase at presentation (primary vs recurrent), type of
institutional series of patients with MLPS/myxoid surgery, tumor size, tumor depth, histotype, patho-
RCLPS and PLPS to investigate prognostic factors and logic grade, margin status, adjuvant treatments, dates
long-term outcomes and to gain a better understand- of neoplastic events, site of distant metastasis (DM),
ing of the possible role of new therapeutic approaches. and death or last follow-up.
Radiotherapy was delivered as an adjunct in 128
patients (39%; 84 patients with primary tumors and
MATERIALS AND METHODS 44 patients with recurrences). The indication for
Between January 1980 and December 2005, 3025 radiation therapy was verified both by the operating
consecutive patients with soft tissue sarcomas under- surgeon and by the radiation oncologist when, based
went surgery with the intent of eradicating disease at on clinical grounds, there was a greater risk of recur-
the National Institute for the Study and Cure of rence. However, no prospectively selected criteria
Tumors in Milan, Italy. In this series, 329 patients were used to this end. External beam radiation was
had a diagnosis of localized MLPS/RCLPS or PLPS. used in all such patients, and doses ranged from 45
Tumors were located mainly at the extremities gray (Gy) to 65 Gy (median, 57 Gy).
and girdles. Retroperitoneal or other abdominal Chemotherapy was received by 61 patients
(intra-abdominal and pelvic) locations were grouped (19%), including 39 patients with primary tumors
together. Paraspinal, chest, and abdominal wall were and 22 patients with recurrent tumors, at the discre-
labeled as trunk locations; whereas the exceptional tion of the multidisciplinary Soft Tissue Sarcoma
head and neck locations were left separate. Tumor Group at our institution or as part of clinical trials.
depth was defined in relation to the investing fascia. Antracycline-based regimens were used, in most
All retroperitoneal/other abdominal locations were patients with ifosfamide.
considered deep. The median follow-up duration for the entire
With regard to the histotypes, over the years, all group, as of June 2006, was 119 months (interquartile
tumors were reviewed by at least 2 experienced range, 70–153 months). Ten patients (3%) were lost
pathologists from our institution. The diagnosis was to follow-up before 10 years. The current study was
confirmed molecularly whenever frozen material was approved by our Institutional Review Board.
available. RCLPS was defined by the presence >5%
round cells. Statistical Methods
The French Federation of Cancer Centers grading The endpoints of this study were disease-specific
system10 was used to assign tumor grade in survival (DSS), local recurrence (LR) incidence, and
untreated primary tumors. Grading for recurrent DM incidence. According to the clinical literature,
tumors was performed on the slides from the pri- curves relating to these endpoints were estimated
mary, untreated tumor seen in consultation. within each subgroup determined by the status of
An attempt to perform a complete resection with disease at presentation to our institution (primary or
negative margins was performed in all patients. For recurrent tumor). DSS curves were calculated by
tumors of the extremities, the surgical procedure was using the Kaplan-Meier method and were compared
similar to what was performed normally for all others by using log-rank tests.
soft tissue sarcomas. Eleven of 273 patients underwent Crude cumulative incidence (CCI) curves for LR
a major amputation. In retroperitoneal, trunk, head and DM were calculated in a competing-risks frame-
and neck locations, wide margins were achievable work, as described by Marubini and Valsecchi,11 and
only rarely. No patients died from surgical complica- comparisons between curves were performed using
tions. Margins were defined at the time of pathologic of the Gray test.12 The time to occurrence of any
assessment by a dedicated pathologist. The surgical event was computed from the date of surgery at our
specimen always was examined in the presence of the institution to the date when the event first was
2524 CANCER June 15, 2007 / Volume 109 / Number 12
recorded or was censored at the date of last follow- to disease-specific death varied from 5 months to
up assessment in event-free patients. 219 months (median for those who died, 54 months);
In the analysis of cause-specific survival, deaths for patients with RCLPS, the time to disease-specific
caused by conditions unrelated to sarcoma were cen- death varied from 5 months to 127 months (median
sored. For the analysis of disease recurrence patterns, for those who died, 27 months); and for the patients
LRs, distant recurrences, and deaths in recurrence- with PLPS, the time to disease-specific death varied
free patients, whichever occurred first, were regarded from 6 months to 195 months (median for those
as competing events. Concomitant LRs and DMs who died, 49 months).
were included in the estimation of the CCI curves as The DSS rates were 83% at 5 years and 75% at 10
DM. years. In the patients who had primary tumors, the
Multivariable analyses of DSS, LR-free survival, DSS rates were 90% and 87% at 5 years and 10 years,
and DM-free survival were based on cause-specific respectively, whereas, in the patients who had recur-
hazards and, thus, were performed using Cox multi- rent tumors, the DSS rates were 72% and 56%,
ple-regression models. Putative prognostic factors respectively, with a significant difference observed
that were included in the models were presentation between the 2 groups (P < .001).
(primary or recurrent), tumor site (extremities or Among the patients who had primary tumors,
other), tumor size (5 cm or >5 cm), tumor grade the 5-year and 10-year DSS rates were 93% and 92%,
(Grade I, II, or III), tumor depth (deep or superficial), respectively, for the MLPS group; 87% and 77%,
margin status (negative or positive), chemotherapy respectively, for the RCLPS group; and 81% and 81%,
(yes or no), radiotherapy (yes or no), and histologic respectively, for the PLPS group. No significant differ-
subtype (MLPS, RCLPS, or PLPS). Because of their ences were observed between the 3 groups (Fig. 1A).
strong correlation, grade and histologic subtype had Among the patients who had recurrent tumors,
to be entered separately into the Cox models. The the 5-year and 10-year DSS rates were 80% and 61%,
effects for the remaining factors were estimated from respectively, for the MLPS group; 59% and 47%,
the models that included histology, the effect of which respectively, for the RCLPS group; and 64% and 53%,
was slightly stronger than that observed for grade. respectively, for the PLPS group. No significant differ-
The distribution of patient and disease charac- ences were observed between the 3 groups (Fig. 1B).
teristics was compared between patients with pri- At multivariable analysis, phase at presentation,
mary tumors and patients with recurrent tumors by tumor size >10 cm, and surgical margins were inde-
means of the Wilcoxon rank-sum test and the Fisher pendent prognostic factors for DSS (Table 2). With
exact test for continuous and categorical variables, regard to histology, there was weak evidence only for
respectively. RCLPS compared with MLPS (P ¼ .09), although
We used SAS software (SAS Institute Inc., Cary, RCLPS and PLPS had a significantly greater propor-
NC) and the S-Plus (StatSci, MathSoft, Seattle, WA) tion of DM in both subgroups of patients (primary
Design Library (available at URL: http://lib.stat. tumors and recurrent tumors), as detailed below. Tu-
cmu.edu) and Cmprsk Library (available at URL: mor grade was an independent prognostic factor if it
http://biowww.dfci.harvard.edu/gray/) to perform was considered instead of histology (P < .05).
the modeling and statistical calculations. For each
test, results were considered statistically significant Local Recurrence
whenever a 2-sided P value <.05 (5%) was achieved. Eighty-six patients developed LR after they under-
went surgery at our institute, including 75 first events
(33 patients with primary tumors and 42 patients
RESULTS with LR at presentation). The time to first LR in the
Clinical characteristics are summarized in Table 1. whole series varied from 2 months to 151 months
(median for those who developed a recurrence, 21
Disease-Specific Survival months). For patients with MLPS, the time to first LR
Sixty-nine patients died of disease (24 of 214 patients varied from 3 months to 151 months (median for
with primary tumors and 45 of 115 patients with those who developed a recurrence, 25 months); for
recurrent tumors), 27 patients died of intercurrent patients with RCLPS, the time to first LR varied from
disease other than sarcoma, and the remaining 233 4 months to 86 months (median for those who devel-
patients were alive at the last follow-up. The time to oped a recurrence, 15 months); and, for patients with
disease-specific death in the entire series varied from PLPS, the time to first LR varied from 2 months to 69
5 months to 219 months (median for those who months (median for those who developed a recur-
died, 48 months). For patients with MLPS, the time rence, 14 months).
Prognostic Factors of MLPS and RCLPS/Fiore et al. 2525
TABLE 1
Main Patient and Disease Characteristics According to the Phase of Disease
IQ range indicates interquartile range; MLPS, myxoid liposarcoma; RCLPS, round cell liposarcoma; PLPS, pleomorphic liposarcoma; FNCLCC, French Federation
of Cancer Centers; RT, radiotherapy; CT, chemotherapy.
* There was 1 missing value in the primary group.
The overall CCI rates for LR were 21.7% at 5 years nificant difference was observed between the 3
and 25.2% at 10 years. In the patients who had pri- groups (Fig. 2A).
mary tumors, the LR CCI rates were 13.2% and 17.5% Among the patients who had recurrent tumors,
at 5 years and at 10 years, respectively; whereas, in the 5-year and 10-year LR CCI rate was 35.3% and
the patients who had recurrent tumors, the respec- 37.1%, respectively, for the MLPS group; 31.3% and
tive rates were 37.1% and 39.2%, and a significant 31.3%, respectively, for the RCLPS group; and 47.8%
difference was obselrved between the 2 groups (P < and 52.2%, respectively, for the PLPS group. No sig-
.001). nificant difference was observed between the 3
Among the patients who had primary tumors, groups (Fig. 2B).
the 5-year and 10-year LR CCI rate was 11.3% and In the multivariable analysis, phase at presenta-
15.9%, respectively, for the MLPS group; 13.7% and tion, tumor site, and radiotherapy were independent
21.9%, respectively, for the RCLPS group; and 19.4% prognostic factors for LR-free survival (a borderline P
and 19.4%, respectively, for the PLPS group. No sig- value of .05 was achieved for radiotherapy), as shown
2526 CANCER June 15, 2007 / Volume 109 / Number 12
TABLE 2
Hazards Ratio Estimates With 95% Confidence Intervals and
P Values From the Cox Proportional Hazards Model on
Disease-specific Survival
HR indicates hazards ratio; 95% CI, 95% confidence interval; RT, radiotherapy; CT, chemotherapy;
RCLPS, round cell liposarcoma; MLPS, myxoid liposarcoma; PLPS, pleomorphic liposarcoma;
FNCLCC, French Federation of Cancer Centers.
TABLE 3
Hazards Ratio Estimates With 95% Confidence Intervals and
P Values From the Cox Proportional Hazards Model on Local
Recurrence-free Survival
HR indicates hazards ratio; 95% CI, 95% confidence interval; RCLPS, round cell liposarcoma; MLPS,
myxoid liposarcoma; PLPS, pleomorphic liposarcoma; RT, radiotherapy; CT, chemotherapy, FNCLCC,
French Federation of Cancer Centers.
TABLE 4
Hazards Ratio Estimates With 95% Confidence Intervals and P Values
From the Cox Proportional Hazards Model on Metastasis-free Survival
HR indicates hazards ratio; 95% CI, 95% confidence interval; RCLPS, round cell liposarcoma; MLPS,
myxoid liposarcoma; PLPS, pleomorphic liposarcoma; RT, radiotherapy; CT, chemotherapy; FNCLCC,
French Federation of Cancer Centers.
TABLE 5
Extrapulmonary Metastatic Pattern According to Histological Subtype*
MLPS indicates myxoid liposarcoma; RCLPS, round cell liposarcoma; PLPS, pleomorphic liposarcoma.
* The number of localizations at single sites (primary/recurrent cases) in 43 patients with extrapulmonary metastases.
TABLE 6
Main Series of Patients With Myxoid Liposarcoma
MLPS indicates myxoid liposarcoma; RCLPS, round cell liposarcoma; LR, local recurrence; DM, distant metastases; M, metastases; DSS, disease—specific survival; NS, not stated.
* Overall survival.
y
Only MLPS and RCLPS.
Patients who had positive surgical margins bore control rates compared with patients who had
a 2-fold greater risk of dying from their disease and a tumors located at other sites. Tumors located in the
1.7-fold greater risk of developing an LR compared trunk bore a nearly 3-fold greater risk of recurring
with patients who did not have positive surgical mar- locally compared with tumors located in the extremi-
gins. The same difference was not observed for ties. This difference in local outcome may reflect the
patients who had metastases. Therefore, patients impact of tumor location on the ability to perform
who had positive surgical margins had a greater risk resection with adequate surgical margins, as reported
of dying mainly from locoregional disease independ- in many series on patients with retroperitoneal and
ent of tumor location, as reported in other series of trunk soft tissue sarcoma,22,23 who also had a nearly
patients with softtissue sarcoma.17–21 2-fold greater risk of dying of their disease, mainly
Two more factors had an affect on recurrence- from locoregional failure. Patients who received
free survival in our series: tumor location and the radiation therapy had approximately 50% of the risk
administration of radiation therapy. Patients who had of developing an LR compared with patients who did
tumors of the extremities had more favorable local not receive radiation therapy.
2530 CANCER June 15, 2007 / Volume 109 / Number 12
With regard to the histologic subtype, a major soft tissue simulators: recognition of new variants and dif-
difference was observed for the development of DM. ferential diagnosis. Virchows Arch. 2001;439:141–151.
6. De Cecco L, Gariboldi M, Reid JF, et al. Gene expression
Patients with primary MLPS had a 10-year incidence
profile identifies a rare epithelioid variant case of pleo-
of DM of 5%, whereas the incidence of DM was 22% morphic liposarcoma carrying FUS-CHOP transcript. Histo-
and 21% for patients with RCLPS and patients with pathology. 2005;46:334–341.
PLPS, respectively. Only minor differences in 10-year 7. Pitson G, Robinson P, Wilke D, et al. Radiation response:
survival were observed, reflecting the indolent course an additional unique signature of myxoid liposarcoma. Int
J Radiat Oncol Biol Phys. 2004;60:522–526.
of the disease even in the metastatic setting and the
8. Jones RL, Fisher C, Al Muderis O, et al. Differential sensi-
relative greater sensitivity of this disease to chemo- tivity of liposarcoma subtypes to chemotherapy. Eur J Can-
therapy.8,9 Furthermore, surgical management of cer. 2005;41:2853–2860.
extrapulmonary metastases often was carried out, 9. Grosso F, Demetri GD, Blay JY, et al. Patterns of tumor
although it sometimes was more challenging in response to trabectedin (ET743) in myxoid liposarcomas.
2006 ASCO Annual Meeting Proceedings. J Clin Oncol.
patients with lung nodules, especially with regard to
2006;24(suppl 18):18S. Abstract 9511.
particular anatomic sites (eg, abdomen, mediasti- 10. Trojani M, Contesso G, Coindre JM, et al. Soft tissue sarco-
num, and heart/pericaridium) (Table 5).24,25 mas of adults: study of pathological prognostic variables
In conclusion, we observed that patients with and definition of a histopathologic grading system. Int J
MLPS, RCLPS, and PLPS who had a confirmed have Cancer. 1984;33:37–42.
11. Marubini E, Valsecchi MG. Analysing Survival Data for
an indolent course, as reported previously using
Clinical Trials and Observational Studies. Chichester: John
predicting tools and based on large case series Wiley & Sons; 1995.
analyses,26–28 Presentation with either primary or re- 12. Gray RJ. A class of k-sample tests for comparing the cumu-
current disease, tumor size, and tumor grade are of lative incidence of a competing risk. Ann Stat. 1988;16:
major importance. Complete tumor resection with 1141–1154.
13. Kilpatrick SE, Doyon J, Choong PF, et al. The clinicopatho-
negative margins should be the objective of surgery
logic spectrum of myxoid and round cell liposarcoma: a
with an evident impact on local control. Adjuvant study of 95 cases. Cancer. 1996;77:1450–1458.
radiation therapy should be delivered to improve 14. Antonescu CR, Tschernyavsky SJ, Decuseara R, et al. Prog-
local control. High-grade and large RCLPS tumros nostic impact of P53 status, TLS-CHOP fusion transcript
have a more aggressive behavior; thus, patients with structure, and histological grade in myxoid liposarcoma: a
molecular and clinicopathologic study of 82 cases. Clin
these tumors may be considered for possible new ad-
Cancer Res. 2001;7:3977–3987.
juvant medical treatments. The cutoff level for 15. ten Heuvel SE, Hoekstra HJ, van Ginkel RJ, Bastiaannet E,
tumors of a 5% round cell component singles out a Suurmeijer AJ. Clinicopathologic prognostic factors in myx-
group of diseases with a broad spectrum of aggres- oid liposarcoma: a retrospective study of 49 patients with
siveness. Further studies will be needed to determine long-term follow-up. Ann Surg Oncol. 2006;14:222–229.
16. Moore Dalal K, Kattan MW, Antonescu CR, et al. Subtype
whether more precise tools may be obtained by stra-
specific prognostic nomogram for patients with primary
tifying the round cell cutoff level further or by exam- liposarcoma of the retroperitoneum, extremity, or trunk.
ining the different molecular subtypes, allowing a Ann Surg. 2006;244:381–391.
better selection of high-risk patients as candidate for 17. Enneking WP, Spainer SS, Goodman MA. A system for the
newly available agents. surgical staging of musculoskeletal sarcoma. Clin Orthop.
1980;153:106–120.
18. Trovik CS, Bauer HCF, Alvegard TA, et al. Surgical margins,
REFERENCES local recurrence and metastasis in soft tissue sarcomas:
1. Weiss SW, Goldblum JR. Enzinger and Weiss’s Soft Tissue 559 surgically-treated patients from the Scandinavian Sar-
Tumours. 4th ed. St. Louis, Mo: Mosby, 2001:641–693. coma Group Register. Eur J Cancer. 2000;36:710–716.
2. Smith TA, Easley KA, Goldblum JR. Myxoid/round cell lipo- 19. Stojadinovic A, Leung DHY, Hoos A, et al. Analysis of the
sarcoma of the extremities. A clinicopathologic study of 29 prognostic significance of microscopic margins in 2084
cases with particular attention to extent of round cell lipo- localized primary adult soft tissue sarcomas. Ann Surg.
sarcoma. Am J Surg Pathol. 1996;20:171–180. 2002;235:424–433.
3. Knight JC, Renwick PJ, Clin PD, et al. Translocation 20. Zagars GK, Ballo MT, Pisters PWT, et al. Prognostic factors
t(12;16)(q13;p11) in myxoid liposarcoma and round cell for patients with localized soft tissue sarcoma treated with
liposarcoma: molecular and cytogenetic analysis. Cancer conservation surgery and radiation therapy. An analysis of
Res. 1995;55:24–27. 1225 patients. Cancer. 2003;97:2530–2543.
4. Mentzel T, Pedentour F. Pleomorphic liposarcoma. In: 21. Gronchi A, Casali PG, Mariani L, et al. Status of surgical
Fletcher CDM, Krishan Unni K, Mertens F, eds. World Health margins and prognosis in adult soft tissue sarcomas of the
Organization Classification of Tumours. Pathology and extremities: a series of patients treated at a single institu-
Genetics of Tumours of Soft Tissue and Bone. Lyon: Interna- tion. J Clin Oncol. 2005;1:96–104.
tional Agency for Research on Cancer Press; 2002:44–45. 22. Singer S, Corson JM, Demetri GD, et al. Prognostic factors
5. Meis-Kindblom JM, Sjogren H, Kindblom LG, et al. Cytoge- predictive of survival for truncal and retroperitoneal soft
netic and molecular genetic analysis of liposarcoma and its tissue sarcoma. Ann Surg. 1995;221:185–195.
Prognostic Factors of MLPS and RCLPS/Fiore et al. 2531
23. Gronchi A, Casali GP, Fiore M, et al. Retroperitoneal 26. Kattan MW, Leung DHY, Brennan MF. Postoperative nomo-
soft tissue sarcoma: patterns of recurrence in 167 patients gram for 12-year sarcoma-specific death. J Clin Oncol.
treated at a single institution. Cancer. 2004;100:2448– 2002;20:791–796.
2455. 27. Eilber FC, Brennan MF, Eilber FR, et al. Validation of the
24. Pearlstone DB, Pisters PW, Bold RJ, et al. Patterns of recur- postoperative nomogram for 12-year sarcoma-specific
rence in extremity liposarcoma: implication for staging and mortality. Cancer. 2004;101:2270–2275.
follow-up. Cancer. 1999;85:85–92. 28. Mariani L, Miceli R, Kattan MW, et al. Validation and adap-
25. Estourgie SH, Nielsen GP, Ott MJ. Metastatic patterns of ex- tation of a nomogram for predicting the survival of
tremity myxoid liposarcoma and their outcome. J Surg patients with extremity soft tissue sarcoma using a three-
Oncol. 2002;80:89–93. grade system. Cancer. 2005;103:402–408.