Deep-seated, well differentiated lipomatous tumors of the chest wall and extremities

The role of cytogenetics in classification and prognostication

Authors


Abstract

BACKGROUND

Intramuscular lipomas and atypical lipomatous tumors (ALT) are common deep-seated lipomatous tumors of the chest wall and extremities. Distinguishing between these two entities can be difficult based on histologic analysis alone. However, the cytogenetic profiles of ALT and intramuscular lipomas are distinct. Correct classification is important, because aggressive local disease recurrence occurs more frequently in patients with ALT than in patients with intramuscular lipoma. The authors examined their single institutional experience and correlated their classification with clinical features and outcome.

METHODS

In the current study, 106 patients with deep-seated, well differentiated adipose tumors of the chest wall and extremities were classified as having ALT or intramuscular lipoma using a combined approach of histology and cytogenetics, if available. The classification was correlated with clinicopathologic features and follow-up data.

RESULTS

Fifty-five patients were classified as having intramuscular lipoma and 51 were classified as having ALT. Classification did not correlate with age and gender (P = 0.28 and P = 0.96, respectively). Intramuscular lipomas were smaller than ALTs (P < 0.0001), but there was significant overlap between the 2 groups. ALT occurred preferentially in the lower extremity (P < 0.0009). Four percent of patients with intramuscular lipomas and 27% of patients with ALTs developed local disease recurrence (P = 0.0006). Disease recurrence did not correlate with patient age at diagnosis, patient gender, tumor size, and tumor location (P = 0.45, P = 0.26, P = 0.49, and P = 0.28, respectively). Within the subset of patients with ALTs, disease recurrence did not correlate with patient age at diagnosis, patient gender, or tumor location (P = 0.38, P = 0.54, and P = 0.86, respectively).

CONCLUSIONS

Classification of deep-seated, well differentiated lipomatous tumors of the extremities and chest wall using a combined approach of histology and cytogenetics correlated well with biologic behavior/disease recurrence. This combined approach is advocated to better stratify patients for treatment purposes and follow-up. Cancer 2005. © 2004 American Cancer Society.

Liposarcoma is the single most common soft tissue sarcoma of adulthood, accounting for approximately 20% of all soft tissue sarcomas.1 There are three recognized histologic subtypes including well differentiated/dedifferentiated, myxoid and round cell, and pleomorphic liposarcomas. Historically, the hallmark of well differentiated liposarcoma has been the presence of multivacuolated lipoblasts. However, these are often rare or inapparent and they are no longer required to support this diagnosis.2 Cytogenetic analysis of well differentiated liposarcoma typically reveals the presence of giant ring and marker chromosomes composed of the q12–15 region of chromosome 12.3–5

Well differentiated liposarcoma can be subdivided clinically into tumors that occur in the chest wall and extremities and tumors that arise in the abdominal cavity/retroperitoneum and mediastinum. Tumors that arise in the chest wall and extremities have the potential to exhibit locally aggressive behavior, including the ability to recur locally. However, these tumors have no metastatic potential unless they dedifferentiate, which is very rare in the extremities and chest wall. Local disease recurrences invariably require additional local excision, with morbidity associated with disease recurrence and surgery. However, these disease recurrences rarely result in patient death. Well differentiated liposarcomas of the abdominal cavity/retroperitoneum and mediastinum often result in patient death due to repeated local disease recurrence with engulfment of vital structures. Patients ultimately die of the complications of bowel obstruction, renal failure, or sepsis. Based on the differing clinical scenarios, it has been proposed that the term “atypical lipomatous tumor” (ALT) should be used for tumors that arise in the chest wall and extremities, whereas the term “well differentiated liposarcoma” should be retained for tumors that arise in the abdominal cavity/retroperitoneum and mediastinum.1, 2, 6

Not all deep-seated fatty tumors have a propensity for local disease recurrence. Intramuscular lipomas are characterized by a histologic appearance that is similar to normal subcutaneous adipose tissue, except that they are located within skeletal muscle.7–9 In contrast to most cases of the lipoma-like variant of ALTs, there are no atypical stromal cells or lipoblasts. Cytogenetic studies reveal simple translocations or loss of chromosomal material involving the q14–15 region of chromosome 12, paracentric or pericentric inversions of chromosome 12q14–15, aberrations involving 6p21–22, or loss of material from the q12–14 or q22 region of chromosome 13.10, 11 Intramuscular lipomas are believed to have a very low rate of local disease recurrence and no metastatic potential. It is possible to classify correctly most patients with intramuscular lipoma. However, in a few patients, the cytologic atypia is so minimal that it is difficult, if not impossible, to distinguish reliably ALTs from intramuscular lipomas based on histology.

In the current study, we examined our experience with deep-seated, well differentiated lipomatous tumors of the chest wall and extremities. Specifically, we correlated the clinicohistologic and karyotypic features with clinical outcome to better understand the classification and clinical behavior of these neoplasms.

MATERIALS AND METHODS

The current study received approval from the University of Washington (Seattle, WA) human subjects committee. The surgical pathology database at the University of Washington Medical Center includes the pathology records of all patients treated at the medical center from 1984 to the present. The database was searched in August 2002 to identify all patients with lipoma and liposarcoma. One hundred six patients with deep-seated (involving or deep to fascia), well differentiated lipomatous tumors arising in the extremity or chest wall were identified. The clinical records of these patients were reviewed and the date of diagnosis of the first occurrence of disease, age at diagnosis, gender, size of tumor, location of tumor, treatment, disease recurrence, and/or metastasis were recorded. The pathology slides for these patients were reviewed by one of the authors (B.P.R.). Tumors comprising uniform, mature adipose tissue that lacked cytologic atypia or multivaculolated lipoblasts were classified histologically as intramuscular lipomas. Tumors with cytologic atypia, usually in the form of fibrous septae containing atypical stromal cells, or lipoblasts, were classified as ALTs. Typically, one section per centimeter of tumor specimen was examined histologically. Chromosome analyses were carried out on G-banded preparations by trypsin treatment and stained with Wright's stain (Giemsa Trypsin Wright [GTW]) according to a modified protocol published by Seabright.12 Statistical analyses were performed using Stat View for Macintosh (Version 5.0.1, SAS Institute, Cary, NC). Comparisons between clinical characteristics were done using the chi-square test and the t test. Significance was set at P ≤ 0.05.

RESULTS

The clinical findings of the 106 patients diagnosed with deep-seated, well differentiated lipomatous tumors involving the chest wall or extremity are summarized in Table 1. The sample comprised 46 men and 60 women with a median age at diagnosis of 54 years (range, 31–85 years). For 60% of the patients, the thigh/buttock was the most common site of involvement by lipomatous tumors. All tumors were excised marginally without a rim of normal tissue. Neither radiotherapy nor chemotherapy was received by any of the patients. The mean tumor size was 13.2 cm and the median tumor size was 11.2 cm (range, 1.1–35 cm). With a median follow-up of 61 months from the time of diagnosis to the time of analysis, 16 patients had local disease recurrence. Ten of the lesions recurred once, 3 lesions recurred twice, 2 lesions recurred 3 times, and 1 lesion recurred 7 times over a 30-year period. The median time to first disease recurrence was 48 months (range, 13.4–149.3 months). There were no metastases. One 16.5-cm ALT of the calf recurred as dedifferentiated liposarcoma 26 months after the initial excision. The disease recurrence involved the tibia extensively and this patient received a below the knee amputation.

Table 1. Clinical Findings in 106 Patients with Deep-Seated, Well Differentiated Lipomatous Tumors of the Chest Wall and Extremities
CharacteristicsTotalIntramuscular lipomaAtypical lipomatous tumorP value
Gender46 males/60 females24 males/31 females22 males/29 females0.96
Median age (range) (yrs)54 (31–85)54 (31–76)57 (34–85)0.36
Site   0.0009
 Chest wall642 
 Shoulder/upper arm22202 
 Forearm624 
 Thigh/buttocks662739 
 Knee/lower leg624 
Median size (range) (cm)11 (1.1–35)10 (1.1–18.5)18 (2.0–35)0.0001
Disease recurrence16 (13.4–149.3 mos)2 (4%)14 (27%)0.0006

Tumor specimens that comprised uniform, mature adipose tissue that lacked cytologic atypia or multivaculolated lipoblasts were classified histologically as intramuscular lipoma (Fig. 1A). Tumor specimens with cytologic atypia, usually in the form of fibrous septae containing atypical stromal cells, or lipoblasts were classified as ALT (Fig. 1B). Based on histology alone, 57 tumor specimens were classified as intramuscular lipoma and 49 specimens were classified as ALTs. Cytogenetic analysis was performed on 61 tumor specimens. One tumor specimen failed to grow in culture and 18 tumor specimens grew but yielded only normal male or female karyotypes, presumably from the outgrowth of nonneoplastic cells. Abnormal karyotypes were available for 42 tumor specimens (Table 2). Tumor specimens that contained simple translocations or loss of chromosomal material involving the q14–15 region of chromosome 12, paracentric or pericentric inversions of chromosome 12q14–15, aberrations involving 6p21–22, or loss of material from the q12–14 or q22 region of chromosome 13 were classified as intramuscular lipoma (Fig. 2A), whereas tumor specimens that contained ring or giant marker chromosomes were classified as ALT (Fig. 2B). Based on cytogenetic analysis alone, 15 tumor specimens were classified as intramuscular lipoma and 21 were classified as ALT. Six tumor specimens had karyotypes that were not typical of either lipoma or ALT and, therefore, were nondiagnostic (Table 2). For the purposes of the current study, we regarded cytogenetic data as more objective than the histologic interpretation because this is subjective and the differences between intramuscular lipomas and ALTs could be very subtle. Six tumor specimens originally classified as intramuscular lipoma by histology harbored karyotypes characteristic of ALT and were reclassified as ALTs based on these karyotypes. Four tumor specimens originally classified as ALT by histology alone possessed karyotypes typical of lipoma and thus were reclassified as intramuscular lipoma. After reclassification by cytogenetics, when available, 55 specimens were classified as intramuscular lipomas and 51 were classified as ALTs.

Figure 1.

(A) Medium power photomicrograph of a representative intramuscular lipoma showing infiltration of skeletal muscle fibers by a bland appearing proliferation of adipocytes. (B) Medium power photomicrograph of a representative atypical lipomatous tumor characterized by a well differentiated lipomatous neoplasm containing fibrous septae with atypical spindle cells.

Table 2. Cytogenetetic Analysis of 42 Deep-Seated, Well Differentiated Lipomatous Tumors of the Chest Wall and Extremities
Specimen no.Karyotype
Lipoma 
 1346,XY,t(5;12)(q33;q13)[16]
 1646,XY,t(4;12)(q22;q14 or 15)
 1846,XY,t(1;12)(p32-34;q13)[15]/46,XY[5]
 2546,XY,inv(12)(p13;q21)[18]/46,XY[2]
 2845,XY,t(3;12)(q27-28;q13-14),−13[3]/46,XY,idem,+mar[17]
 4646,XY,del(2)(p21p23),t(9;12;12)(q34;p12;q21)[17]/46,XY[3]
 5146,XY,t(5;12)(q33;q13)[20]
 5246,XY,del(12)(q12q15),t(13;14;13)(?q32;q31;q14)[13]/46,XY[7]
 5346,XX,t(3;12)(q27-28;q13-14)[19]/46,XX[1]
 5446-92,XXYY,t(13;13)(q14;q34)[16]/46,XY[4]
 8146,XY,add(1)(p?22),der(8)t(7;8)(p11;p11),add(11)(q?13)add(12)(q15 or 21)[6]/46,XY,add(1)(p?22),?inv(2)(p?25q?14),der(8),t(7;8)(p11;p11),add(11)(q?13),add(12)(q15 or 21)[4]/46,XY[10]
 8646,XY,t(6;16;12)(q21;p11;q13),del(13)(?q21)[20]
 8746,XY,t(3;5;12)(q21;q31;q21),del(13)(q12-21.2)[16]
 9246,XX,t(12;15)(q13;q13)[7]/46,XX[13]
 9946,XY,inv(12)(p12;q14)[20]/46,XY[2]
Atypical lipomatous tumors 
 245-48,XX,+14[3],+r[6][cp8]/93-94,XXXX,+r[2]/46,XX[10]
 497,XXXX,−5,−5,+6,−8,add(10)(p15),add(11)(p15)x3,+13,−15,−16,−16,del(17)(p11.2),add(18)(q23),+19,+r,+8markers/[1]/48,XX,+del(1)(q32),+9, add(11)(p15),−12,−12,+r,+mar[1]/46,XX,+1,−7,add(11)(q25),−12,−14,−21,−21,+r,+2 markers
 1146-49,XY[20]add(12)?t(?6;?8?12)(?;?;?q13)+r[2],+mar[6][cp20]
 2246,XY,t(3;11)(q25;q13),t(7;11)(q22;q13)[5]/46,idem,−8,+20,+mar[1]/46,XY,t(4;7)(q21;q21)[1]/46,XY,?inv(16)(p11.2q22),?add(17)(q25)[1]/46,XY[13]
 3047,XX,+r[7]/47,XX,+mar[6]/46,XX[5]
 3743-49,XX,dic(10;19)(q26;p13.3),−20,+r1,+r2,+mar[cp6]/46,XX[14]
 3846,XY,trp(12)(q14q24.3)[20]
 4246-48,XY,dic(2,3)(q35;p25)[2],add(8)(q24)[2],+9[3],dic(15;22)(p11;q13)[4],+r[3][cp6]/46,XY[14]
 5040,dic(X;9)(q28;p24),−T,der(3)t(1;3)(?q32;p26),−9del(9)(p23),−11,der(12)t(?;12)(?;q22),dic(13;14)(q34;p13),−16,−19,−19,−21,+r,+r[1]/46,XY[1]
 5745,XY,t(1;5)(p13;q?32),t(2;8) (p23;q24.3),add(17)(p13)[17]/45-47,XY,+X,der(1)t(1;5)(p13q?32),t(2;8)(p23;q24),add(5)(q?31),−12,−14,add(17)(p13),−18,+mar1,+mar2,+mar3,+mar4,+2-4mar,+1-3r[3]
 5842-48,X,−Y[8],del(2)(q33)[2],−13[4],−22[3],+r[4],+mar[4][cp9]
 5947,XY,+r[6]/46,xy[14]
 6349,XY,−5,+11,+12,−21,−22,+r,+r,+r/48,XY,−12,+r,+mar,+mar/46,XY,−19,+mar/46,XY
 6539-47,XX,t(3;11)(q27;q14),del(6)(q23),+r
 6846-47,XX,add(2)(p25)[3],−6[4],+mar[18][cp19]/46,xx[1]
 7944-50,XX,add(7)(p22)[11],−22[5],+1-4 dmin[4],+1-7mar[10],+r[18][cp18]/46,XX[2]
 8447,XY,+r[10]/46,XY,[11]
 8947-49,XY,+1-2 r[6]/46,XY,t(1;1)(q42;q21)[2]/46,XY[12]
 9047,XX,+r[2]/46,XX,[18]
 9143-51,XY,+3[2],−5[2],−9[2],+13[2],+19[2],−20[2],+22[3],−22[2],+r[6][cp9]
 9649,XX,+r,+mar1,+mar2[9]/49,XX,+r,+mar1,+mar2,min[4]/49,XX,+1,−4[1]1,−21[1],_r,+mar1,+mar2[cp2]/46,XX[5]
Nondiagnostic 
 14[46,XX,add(11),(q12-13)
 2646,XX,t(1;11)(q12;q13)[16]/46XX[4]
 3547,XX,+X[5]/46,XX[15]
 4746,XX,t(1;10)(q32;p11.2)[4]/46,XX[16]
 7046,XX,del(10)(q24)[20]
 7792,XXYY,+?17,−22,+idic(22)(q13)[2]/92,XXYY[4]/46,XY[13]
Figure 2.

(A) Karyotype of a lipoma with a reciprocal translocation between the long arms of chromosomes 3 and 12 (see arrows). (B) Karyotype of an atypical lipomatous tumor with a large ring chromosome (see arrow).

After final classification as intramuscular lipoma and ALT, a comparison of several parameters was performed (Table 1). A significant difference in age or gender was not observed between the 2 histologic groups (P = 0.28 and P = 0.96, respectively). Intramuscular lipomas occurred in 24 males and 31 females with a median age of 54 years (range, 31–76 years), whereas ALTs arose in 22 males and 29 females with a median age of 57 years (range, 34–85 years). Intramuscular lipomas had an equal predilection for the upper and lower extremity, whereas a high proportion of tumor specimens classified as ALTs occurred in the lower extremities (P < 0.0009). Intramuscular lipomas were found in the chest wall (n = 4 patients), arm (n = 20 patients), forearm (n = 2 patients), thigh (n = 27 patients), and leg (n = 2 patients), whereas ALTs were identified in the chest wall (n = 2 patients), arm (n = 2 patients), forearm (n = 4 patients), thigh (n = 39 patients), and leg (n = 4 patients). Intramuscular lipomas were smaller than ALTs (P < 0.0001), but there was considerable overlap between the 2 groups. Intramuscular lipomas had a median size of 10.0 cm (range, 1.1–18.5 cm), while ALTs had a median sized of 18.0 cm (range, 2.0–35 cm). A statistically significant association between histologic classification and disease recurrence was identified. Of 55 patients with intramuscular lipomas, 2 (4%) developed local disease recurrence, whereas 14 of 51 (27%) patients with ALT developed local disease recurrence (P = 0.0006). Review of the histology of the two patients with intramuscular lipomas who developed disease recurrence revealed that although the tumors diffusely infiltrated normal skeletal muscle, they were composed of very well differentiated adipocytes without atypical stromal cells or lipoblasts. Cytogenetics data were available for only 1 of these patients, a 12-cm thigh mass with a nondiagnostic karyotype that lacked ring or giant marker chromosomes: 92,XXYY, + ?17, -22,+idic(22)(q13)[2]/92,XXYY[4]46,XY[13]. This tumor recurred 70 months after initial excision. The other lesion was a 15-cm calf mass that recurred 25 months after initial excision. Disease recurrence did not correlate with patient age at diagnosis, gender of the patient, tumor size, or location of the tumor (P = 0.45, P = 0.26, P = 0.49, and P = 0.28, respectively). Within the subset of patients with ALTs, disease recurrence did not correlate with patient age at diagnosis, gender of the patient, or tumor location (P = 0.38, P = 0.54, and P = 0.86, respectively). Larger ALTs were more likely than smaller lesions to recur after resection (P = 0.03). In 10 of the 14 patients with recurrent ALTs, the primary neoplasm was > 10 cm.

DISCUSSION

ALTs and intramuscular lipomas are the most common lipomatous neoplasms of the deep soft tissues of the chest wall and extremities. Although it is usually possible to distinguish them from each other using standard diagnostic criteria, there is histologic overlap between these two entities. Some examples of the lipoma-like variants of ALTs lack cytologic atypia and lipoblasts, and are indistinguishable from intramuscular lipomas. This distinction is important, because their rate of local disease recurrence is very different. In the current study, 4% of patients with intramuscular lipomas had disease recurrence, whereas 27% of patients with ALTs had disease recurrence after marginal resection. There was a statistically significant difference in size between intramuscular lipomas and ALTs (P < 0.0001). However, there is great overlap in size between these two groups so size alone cannot reliably distinguish ALTs from intramuscular lipomas.

Cytogenetics can be helpful in distinguishing intramuscular lipomas from ALTs as lipomas tend to harbor simple translocations or loss of chromosomal material involving the q14–15 region of chromosome 12, paracentric or pericentric inversions of chromosome 12q14–15, aberrations involving 6p21–22, or loss of material from the q12–14 or q22 region of chromosome 13.4, 5, 11 The rearrangements involving 12q14–15 and 6p21–22 lead to overexpression of HMGIC and HMGIY, respectively.13–16 The proteins encoded by these genes are high mobility group proteins that are involved in determining chromosomal structure and are known to affect gene expression globally. ALTs have characteristic giant marker and ring chromosomes.4, 5 The giant marker and ring chromosomes are composed of amplicons of the 12q13–15 region, which contains many potential oncogenes including MDM2, CDK4, HMGA2 (also known as HMGIC), SAS, GLI, DDIT3 (also known as CHOP), OS4, and OS9.17 Investigations are under way to determine which subset of these genes is responsible for driving the neoplastic process.

In the current study, 6 of 57 patients (10%) originally classified as having intramuscular lipoma were reclassified as having ALT based on the observation that their karyotypes revealed giant marker or ring chromosomes (Table 2). Indeed, two of these patients developed local disease recurrence and had clinical characteristics suggestive of the diagnosis of ALT. Five of the six patients had large thigh masses, which, from the data in the current series, had a higher likelihood of aggressive behavior. The other patient had a 7.5-cm forearm mass. Furthermore, 4 of 49 (8%) patients with tumors that histologically appeared to be ALT were reclassified as lipomas based on their karyotypes. These four reclassified patients did not develop disease recurrence. Three of the reclassified patients had lipomas that occurred in the thigh and one had a lipoma that arose in the shoulder. Histologic reexamination of these four patients revealed minimal but definite cytologic atypia, with no fibrous septae containing atypical cells or lipoblasts. It appears that based on histology alone, it is possible to misdiagnose intramuscular lipoma as ALT and vice versa.

Not all karyotypes in the current study provided useful information. Six of the karyotypes were nondiagnostic, i.e., they did not harbor either a translocation or chromosomal rearrangement associated with lipomas or the giant marker or ring chromosomes associated with ALTs. These included two patients who were classified as having lipoma and four who were classified as having ALT by histology. We used the histologic classification to classify these patients for the purposes of statistical evaluation. One of the specimens classified as lipoma by histology contained only an extra X chromosome whereas the other contained a near tetraploid karyotype. The four specimens classified as ALT had relatively simple karyotypes with the addition of material to the long arm of chromosome 11 in 1 and t(1;11)(q12;q13), t(1;10)(q32;p11.2), and del(10)(q24) in the others. None of these karyotypes have been reported to be associated with either lipomas or ALTs and they are all different from each other. Aberrations involving chromosome 11q are associated with hibernomas but reexamination of the pathology in the specimen with additional material on the long arm of chromosome 11 revealed a lipoma-like ALT without the histologic features of hibernoma.18

Because only two patients had recurrence of lipomas in the current study, we reexamined their specimens to determine whether there was anything that we could learn from them. Review of the histology revealed very well differentiated lipomatous tumors without atypical stromal cells or lipoblasts. It is noteworthy that the tumors diffusely infiltrated normal skeletal muscle at the periphery of the lesion. Fletcher and Martin-Bates8 also reported that infiltration, in contrast to circumscribed margins, was associated with recurrence of intramuscular lipomas. Cytogenetics were performed on one of the two lipomas that recurred in this study, and were not diagnostic of lipoma or ALT. Based on the very low rate (4%) of local disease recurrence in the group we diagnosed as intramuscular lipoma, we believe that our criteria for distinguishing intramuscular lipomas from ALTs are valid, and we continue to use both histologic evaluation as well as karyotype in the classification of deep-seated lipomatous tumors. When there is disagreement between the histologic appearance and the karyotype, we classify the tumor as ALT or lipoma based on the karyotype.

The anatomic distribution between intramuscular lipoma and ALT differed in our study. Intramuscular lipomas occurred more frequently in the upper extremities and chest wall than ALTs and comprised 76% of the tumors in these anatomic locations. In our series, 47% of the intramuscular lipomas arose in the upper extremity or chest wall, and this predilection of intramuscular lipomas has been reported previously by Fletcher and Martin-Bates.8 In that study, 62% of the intramuscular lipomas arose in the trunk or upper extremity.8 In contrast, only 16% of the ALTs (well differentiated liposarcomas) arose in the upper extremities or chest wall, which is consistent with a previous report.19

The disease recurrence rate of intramuscular lipoma is believed to be very low. In our study, the disease recurrence rate was 4% and is at the low end of the 3–62.5% recurrence rate that has been reported in the literature.8, 9, 20 Studies that reported a high rate of local disease recurrence were performed at a time when lipoblasts were required for the diagnosis of ALTs. Therefore, these studies likely misclassified many patients with the lipoma-like variant of ALTs/well differentiated liposarcoma as having intramuscular lipoma. It is also possible that the true disease recurrence rate of intramuscular lipoma approaches zero and that the two intramuscular lipomas that recurred in our study are actually unrecognized ALTs. However, without a diagnostic karyotype or histologic evidence of ALT, we believe that it is not appropriate or feasible to classify these tumors as ALTs.

The disease recurrence rate of the ALTs in the current study (27%) falls into the middle range of what has been reported in other studies. The reported rates of local disease recurrence in other studies ranges from 5% to 52%.19, 21–23 The median follow-up period in our series was 51 months of patients with ALT, whereas the median time to local disease recurrence was 48 months. Therefore, the estimate of local disease recurrence in our series likely underestimates the true rate of local disease recurrence, which will likely increase over time. In a recent study from Memorial Sloan Kettering, Kooby et al.24 reported that disease recurrence developed in patients with the sclerosing subtype of ALTs, which constituted a large percentage of their cases. In the current study, only 2 patients (4%) had the sclerosing subtype of ALTs and neither of these patients had disease recurrence. In our study, all recurrent ALTs were of the lipoma-like subtype.

Well differentiated liposarcoma of the groin and retroperitoneum transforms to dedifferentiated liposarcoma with a relatively high frequency compared with ALTs of the trunk and extremities. We summarized the results of several studies that investigated dedifferentiation and found that only 1% (8 of 258) of ALTs of the extremities underwent dedifferentiation, whereas 26% (22 of 86) of well differentiated liposarcomas of the retroperitoneum underwent dedifferentiation.2, 6, 19, 22–25 In the current study, we identified a single case (2%) that underwent dedifferentiation. Therefore, our results are in line with the results from other studies. Given the extremely low rate of dedifferentiation and the lack of mortality associated with ALTs of the extremities and trunk, it seems reasonable to us to use the term ALT when referring to these lesions in the extremities and trunk. This is in contrast to retroperitoneal lesions, which are associated with much higher rates of transformation to dedifferentiated liposarcoma and mortality due to an inability to resect lesions completely and are, therefore, best designated as well differentiated liposarcoma.

As we have demonstrated in the current study, recent advances in our understanding of the classification of deep-seated, well differentiated lipomatous tumors have allowed us to define which tumors have a higher likelihood of local disease recurrence. However, it remains to be determined whether this is meaningful from a clinical standpoint. The current practice at the University of Washington is to perform marginal resection of all such tumors, regardless of the potential for local disease recurrence. The justification for this practice is that wide resection of all ALTs is associated with more morbidity than reexcision of isolated tumors that recur. ALTs tend to be large at the time of diagnosis, and often abut important nerves and blood vessels. In a recent paper by the group at Memorial Sloan Kettering, a local disease recurrence rate of only 5.4% was reported, but the authors stressed that their institutional practice is to perform a reexcision of tumors that were incompletely excised.24 In addition, 17 of 91 of their patients received radiotherapy in an attempt to provide additional local control of disease. Because the morbidity and impact on patient quality of life were not evaluated for either this more aggressive approach or the more conservative practice at our institution, it is difficult to fully compare the outcomes of the two studies. However, this would be an interesting and important area for further study. In the meantime, we believe that accurate classification of these tumors is important, and as demonstrated in the current study, cytogenetic studies can provide additional data that aid in the classification of these histologically challenging tumors. We advocate that a combined approach be used to classify these lesions to better stratify patients for treatment purposes and follow-up.

Acknowledgements

The authors thank Christopher D.M. Fletcher, M.D., for his helpful comments regarding the current report.

Ancillary