Günter Köhler, Katja Evert, Marek Zygmunt and Matthias Evert
LMS is a malignant tumor of the smooth musculature. LMS account for 23.9% of all STS that exhibit a particular type of differentiation, and around 40% are of uterine origin (295). They also arise in the ovaries and in the vagina in the vicinity of the vulva, as well as retroperitoneally, predominantly within the ligamentum latum.
Leiomyosarcomas are the most common type of soft tissue sarcoma and occur in the entire genital region.
The new WHO Classification differentiates between spindle-celled, epithelioid and myxoid LMS (220).
In the past, LMS have – almost without exception – been covered together with other uterine sarcomas and malignant mixed tumors, mostly in retrospective or singlearm prospective studies, often with only very small case numbers. The only available randomized studies (223, 249) on adjuvant CHT and postoperative RT also covered all types of uterine sarcoma. While uterine LMS have indeed been included in larger randomized STS studies (185, 303), they only ever played a subordinate role in the analyses performed. The problems pertaining to the data situation are further exacerbated by the fact that the studies on uterine LMS almost always cover tumors in stages I to IV and also include recurrent and metastasized tumors together. Recently, several extensive retrospective and prospective studies have been published that also allow adjuvant-therapeutic conclusions to be drawn for R0 resected LMS. Furthermore, two randomized studies on the application of CHT to advanced and metastasized LMS have recently been published (124, 235). Another problem lies in the fact that LMS can be very difficult to diagnose on the basis of microscopy. Therefore, up to 30% of cellular, myxoid, epithelioid and even regular LM as well as LM with elevated mitotic activity and STUMP are wrongly classified as LMS (101). Overall, this state of affairs severely impedes making critical “evidence-based” assessments and evaluations of previous research.
Despite their relative frequency, exclusive uterine leiomyosarcomas have been the subject of only two randomized studies. While some retrospective analyses have been very extensive, they have been the subject of numerous biases and usually cover all stages of disease. “Evidence-based” therapy recommendations are the exception.
LMS constitutes the most common type of uterine sarcoma, accounting for 25–30%. In Germany, LMS account for 32.2% of all uterine sarcomas when CS are included, and 46.7% when CS are excluded (157). In Norway, LMS account for 68% (4). The annual incidence rate is 0.55/100,000 among white women and 0.92/100,000 for black women (295). In northern Europe, with around 0.4 cases/100,000 women of all age groups, the highest incidence rates could be observed among women aged between 45 and 59 (160). The incidence and prevalence ofLMS do not correlate with those of LM, the latter has a prevalence rate of between 11.7% and 23.6%(68). The median and mean ages of disease onset are 52 and 53.9 years, respectively (min. 31, max. 90 years) (157). Other sources point to a median age of 48–57 years (4, 11, 130, 170, 246, 247, 256). The median age of patients with myxoid and epithelioid LMS corresponds to that of patients with regular LMS (4, 41). Fifty percent of patients are thus postmenopausal. According to DKSM data, 62.5% of women are postmenopausal and 66.1% are aged older than 50 (157).Women with LMS are considerably older than patients with STUMP. One source states a median age of 47.2 years and a mean age of 47.5 years for STUMP (157), while another study states a mean age of 42 years (243). In the literature, the median age of patients with LM ranges from 34–41.5 years (36, 62, 68, 233). The 3,920 LM patients in the DKSM database had a median age of 44 years (157). Besides their somewhat differingmicroscopic criteria, the 7–10 year median age difference between LMS and STUMP patients can be interpreted as suggesting that STUMP probably constitute a precursor for LMS. Disease frequency increases as patients’ age progresses, but is displaced by CS in higher age groups. Only around 5–13% of LMS primarily originate in the cervix (21, 157). Median age (49 years) is lower among patients with cervical LMS (21) than those with corporal LMS. LMS constitute the most common form of primary cervical sarcoma, followed by AS. Only cervical CS has a higher incidence rate (21).
A recently reported decrease in the incidence of LMS cannot be taken at face value and is in fact not true as such. In fact, this development is likely to be a consequence of the fact that, in the course of the past decade, STUMP have come to be removed from LMS statistics increasingly frequently (295). It is not seldom the case that reappraisals reveal tumors that had previously be classified as LMS to in fact be STUMP or LG-ESS with smooth muscle differentiation (4, 329).
LMS develop de novo in the wall of the uterus, independent of LM. The notion that LMS develop from typical LM via malignant transformation has come to be generally rejected (308). Only rarely do findings suggest that an LMS has pathogenetically developed from an LM(199). These LMS are said to have a relatively promising prognosis (316). Judging by their epidemiology, their incidence and their morphologic characteristics, it seems highly likely that STUMP can progress into LMS. In fact, some publications report cases in which areas were found within LMS that consisted of CLM or STUMP, and which exhibited the genetic characteristics of LMS (199, 200). Analyses suggest that p16 might play a role in the pathogenesis of LMS. P16 is often overexpressed in LMS, but usually not in LM. High p16 and p53 expression as well as an elevated Ki67 index are only found in LMS, but not in LM and STUMP. These general features notwithstanding, expression of p16 can be exhibited in CLM and STUMP, especially in metastasized tumors (225) (for details see the section on variants of LM). It is therefore conceivable that LMS can develop from these smooth muscle neoplasms (200).
In contrast to LM, LMS often exhibit chromosomal aberrations and genomic instability. Furthermore, the differing gene expression patterns between LMS on the one hand, and normal myometrium and typical LM on the other hand, strongly suggest that these types of tumors each have their own distinct underlying pathogeneses and molecular pathways (245).
There are data which suggest that a loss of BRCA1 function may play a role in the pathogenesis of LMS (315). Besides other sarcomas, LMS arise more frequently in patients with inherited p53 mutations (Li–Fraumeni syndrome) (126), which serves to underline the causal-pathogenetic role that p53 mutations play in some LMS. When an LMS arises in regular LM, the two components substantially differ in particular in terms of their proliferation index Ki67 (316). The data concerning Her-2/neu overexpression are contradictory. Reported rates of Her-2/neu overexpression in LMS range from 0 to 17%(14, 208, 324). A comprehensive elaboration and presentation of all genetic changes in LMS would go beyond the scope of this monograph and can be read up on in Kobayashi et al. (156).
The frequency to which the medical histories of patients with LMS reveal previous tamoxifen ingestion has been on the increase (321). Interestingly, regular LM appear to arise less frequently under and after tamoxifen therapy (180). LMS growth is generally regarded as being estrogen independent, a view that is supported by the fact that LMS can progress under treatment with GnRH analogues (167). According to an epidemiologic study, HRTwith estrogens and progestins for longer than 5 years is said to elevate the risk of developing LMS (136), though doubts regarding these findings are indeed justified.
In contrast to other genital sarcomas, previous pelvic RT apparently does not play a role in the pathogenesis of LMS (100). Women with hereditary retinoblastoma have a substantial excess risk of uterine LMS (90).
New binding FIGO staging (85, 275) was released for uterine LMS and stromal sarcomas in 2009 (Tab. 2.1.1), replacing the old classification that had been identical to that for EC until the end of 2008. The current staging system also takes the most important prognostic factor “tumor size” into account as well as the fact that LMS can also primarily originate in the cervix,which need not necessarily imply a higher stage. One problem lies in the fact that the results from numerous previous studies cannot be simply transferred or said to apply to the new staging. For example, while stage II in the old staging system included uterus-confined tumors that extend to the cervix, current stage II refers to disease that has spread beyond the uterus yet confined to the pelvis. Stage III has also experienced dramatic adaptation. Tumors that now belong to stage III had previously been categorized as stage IV. Therapeutic recommendations that had in the past applied to stages II and III can thus not simply be adopted for current FIGO and UICC stage II and III.
In the monograph at hand, the old data have been adapted to the modern classification as far as possible.
There are ambitions among some to classify uterine LMS according to the criteria of the “American Joint Committee on Cancer Staging Systems”. The aforementioned classification for uterine LMS is flawed in a number of ways, so that it is currently barely drawn on in practice.
Tab. 2.1.1: Staging for uterine leiomyosarcomas and stromal sarcomas, according to FIGO and UICC 2009.
|T1||I||Tumor limited to uterus|
|T1a||IA||Tumor 5 cm or less in largest dimension|
|T1b||IB||Tumor > 5 cm|
|T2||II||Tumor extends beyond the uterus but is confined to the lesser pelvis|
|T2b||IIB||Involvement of other pelvic tissues/structures|
|T3||III||Tumor invades abdominal tissues|
|T3b||IIIB||> one site|
|N1||IIIC||Metastasis to regional lymph nodes|
|T4||IVA||Tumor invades bladder and/or rectum|
Leiomyosarcoma is the most common type of uterine sarcoma. Patients have a median age of 53 years. LMS develops de novo in the wall of the uterus, and can arise from leiomyoma in exceptional cases. Tamoxifen exposure appears to play a role in this tumor’s pathogenesis. Growth nonetheless appears to be independent of estrogen. Leiomyosarcomas need to be reliably and strictly differentiated from smooth muscle tumors with uncertain malignant potential. But leiomyosarcoma probably develop from these neoplasms.
UICC staging from 2009 applies to uterine leiomyosarcoma and replaces previous staging versions. The decisive change lies in the fact that now, stage I encompasses all tumors that are confined to the uterus including the cervix.
▸ Fig. 2.1.1: (A) the intraoperative aspect of this relatively large leiomyosarcoma reveals serosacovered extension into the abdominal cavity; (B) this infiltration is not visible in the selected axial sonographic image, which does, however, reveal that the uterine structures have practically entirely disappeared leaving only the serosa as a sheath or “shell”; (C) sagittal (left) and axial (right) CT confirmed this finding, revealing that the normal tissue structures had been completely set aside by hypodense tumor masses. These masses are pervaded with contrast-enhancing septa. The infiltration is clearly recognizable in axial CT (arrow). In contrast to benign leiomyoma, no intratumoral calcifications could be found; (D) observing the cut-open uterus reveals that the tumor has entirely consumed the normal uterine structures, the cavum is no longer, the serosa only barely discernible; (E) the strong growth-related pressure emitted by the tumor in (A)–(D) caused the cervix to dilatate, leading to a prolapse from the cervix. This large leiomyosarcoma had a poor prognosis (see Fig. 2.1.19 (A)–(D)); (F) this enormous cervical leiomyosarcoma exhibited the clinical and laparoscopic features of leiomyoma.
LMS are primarily intramurally localized and are thus located within the myometrium in a majority of cases. LMS grow rapidly towards the serosa or the cavum. In doing so, they can displace the cavum or rupture it and subsequently fill it with tumor masses.Myometrial infiltration can be so comprehensive that the myometrium is only covered by the serosa. The tumor can also breach or rupture the serosa and infiltrate the free abdominal cavity (Fig. 2.1.1 (A)–(D), 2.1.7(J)) or prolapse/protrude into the vagina via the cervical canal like a “myoma in statu nascendi” (Fig. 2.1.1 (E), 2.1.4(A)). Such situations can often be clearly reconstructed in sonographic and CT images (Fig. 2.1.1 (B), (C), 2.1.7(I)).
LMS can attain extreme dimensions, and sizes range from 0.7 to 30 cm (Fig. 2.1.1 (F), 2.1.5(A), (B)) (157). Among the 224 LMS covered in the DKSM database, 50% were solitary tumors (see also symptoms, clinical features). When there are multiple “myomas” or in cases of uterus myomatosus, there is the risk that LMS are not recognized as such or do not become visible in sonography (157). Inmost such cases, and in 98% of applicable cases covered in the DKSM database (Fig. 2.1.2 (A)) (157), the LMS constitutes the largest lesion. Even when there are multiple or numerous uterine nodules, LMS practically never accounts for more than one of them (265). Very large LMS can also develop in the cervix and infiltrate the parametria (Fig. 2.1.14).
LMS have a very pleomorphous macroscopic appearance. The outer aspect, or appearance, of solid tumors can be strongly reminiscent of LM (Fig. 2.1.2 (B), (C)). On the cut surface, however, LMS lack the whorled structures that are typical for LM (Fig. 2.1.4 (A)). Cutting causes the LM to bulge and protrude forwards, while LMS tend to remain on a plane with the cut surface. LMS usually have a fleshy, soft consistency and are poorly delineated from the surrounding tissues and structures. What appears to be a capsule is in fact only compressed tumor tissue (Fig. 2.1.1 (D), 2.1.2(D), (E), 2.1.13). The noticeable “softness” of LMS is a feature commonly described in surgery reports, and can cause the LMS to rupture merely upon touching, a risk that is further exacerbated by the absence of a capsule (Fig. 2.1.2 (D), (E)). Therefore, unlike LM, LMS cannot be enucleated. Tumors usually have a pinkish, yellowish, or fish meat-like, grayish-white color. In contrast to LM, LMS usually exhibit necroses, hemorrhages and blood-filled cysts. The latter two features are particularly typical characteristics suggestive of LMS or another uterine sarcoma, are a substantial reason why LMS are so soft (Fig. 2.1.1 (D), 2.1.2(D), 2.1.13). They are often already recognized as “suspicious LM” during surgery, but are not always taken seriously enough. It should be noted that LMS can also have a highly pleomorphous macroscopic appearance (Fig. 2.1.1 (D)).
The necroses present usually have a greenish-yellowish color and thus contribute to the pleomorphous appearance of LMS (Fig. 2.1.13). LMS can, however, also be more or less encapsulated, and can strongly protrude or ooze from the apparent capsule when the latter is perforated/injured. In short, the macroscopic spectrum ranges from a myoma-like appearance, to tumors that are completely necrotically decomposed (features that UUS and CS are known to frequently exhibit) (cf. Chapter 5 and Vol. 2, Chapter 7).
Fig. 2.1.2: (A) supracervical hysterectomy, the leiomyosarcoma constituted the largest of the four clearly visible masses, the three smallest of which were regular leiomyomas that barely differed from the leiomyosarcoma in terms of their gross appearance; (B) this leiomyosarcoma was mistaken for leiomyoma even during supracervical hysterectomy; (C) in contrast to the findings presented in (B), in the tumor in (C) was in fact a leiomyoma that had prolapsed from the cervix; (D) this leiomyosarcoma was extremely soft, exhibited hemorrhage and was located immediately beneath the serosa. It perforated/ruptured immediately during the attempt to lift the uterus out by hand; (E) the corresponding ultrasound image clearly shows that the cystic-necrotic tumor had been covered only by the serosa.
Epithelioid LMS can barely be macroscopically discerned from typical LMS. Myxoid LMS tumors, by contrast, are usually more or less gelatinous, soft, relatively large and poorly circumscribed (Fig. 2.1.4 (A), (B)). Clinical differentiation from myxoid LM is usually no longer possible when observing the cut-open specimen (Fig. 2.1.4 (B)). Myxoid LMS have a median diameter of 8 cm (41).
In terms of their microscopic features, LMS can be categorized into three histologic groups – spindle-celled, epithelioid and myxoid LMS (222). Spindle-celled LMS – the histologic prototype of LMS – is generally highly cellular and consists of intersecting bundles of spindle cells and/or pleomorphous cells with eosinophilic cytoplasm. The strong degree of cellularity does not play a role in categorizing LMS as such. LMS usually exhibit destructive myometrial infiltration. VI is a typical feature, and is the primary cause for the early development of distant metastasization. Such VI is found in at least 30–34% of non-epithelioid or non-myxoid LMS (4, 59). The presence of TCN, mitotic activity and nuclear and cytologic atypia are the three decisive factors for classifying a lesion as LMS (25, 50, 217, 294).
The characteristic TCN constitute the most important diagnostic feature of LMS (Fig. 2.1.3 (A)). They can be found in 55–89% of all LMS (4, 156). TCN are characterized by their typically irregular yet sharp margins to vital tissue. Since their margins are strongly reminiscent of the outlines of continents on a map, they are also referred to as “geographic” necrosis. “Coagulative necrosis” is another synonym. TCN usually arise multifocally. Hyaline or infarction-like necroses and surrounding vital tissues are separated by a transitional zone, while TCN are not. This zone consists of fibrous hyalinized and/or granulated tissue in the sense of potential previous inflammatory or reactive processes. Differentiating between these two types of necrosis can be challenging. Necroses forwhich such a differentiation cannot be reliablymade are referred to as uncertain or questionable TCN. Diagnostically irrelevant hyaline necroses are also observed in up to 61% of LMS (4). It should be noted that, in TCN, vital tumor cells are often still found located around larger vessels.
The number of mitoses, the second diagnostic pillar, is determined via hematoxylin and eosin staining (Fig. 2.1.3 (B)). The mitotic index is determined by the mean number of mitoses per 10 HPF. Measuring at least 30 fields in the “hot spot” (the area with the highest number of mitoses) is said to yield precise results. Atypical mitoses are also often encountered, and can be regarded as a sign of genomic instability. Counting mitotic activity using H&E staining is very laborious and time consuming, and the results can be distorted or falsified by the presence of apoptotic structures or pyknotic nuclei. Consequently, the immunohistochemical marker PHH 3 has seen increased use for measuring mitotic activity. PHH 3 responds specifically to cells in which a mitosis is currently taking place (229, 305). Mitotic count is also measured per 10HPF using this method. PHH 3-IHC is a relatively straightforward procedure, is far less time consuming than H&E staining, and more closely correlates with the Ki67 index (229). Another advantage over H&E staining lies in the fact that PHH 3 does not react with hyperchromatic or pyknotic nuclei, and thus essentially “masks out” necroses and apoptoses.
Since mitoses predominantly occur in „hot spots“, the count should not be performed in a randomly selected area of the tumor, a requirement that punch biopsy does not fulfill.
Significant diffuse or multifocal, moderate-to-severe nuclear and cytologic atypia constitute the third diagnostic pillar (Fig. 2.1.3 (C), (D)). Atypia that are already recognizable as such under × 10 magnification (also referred to as “low power field”) are termed “significant atypia”. Nuclear atypia are characterized by hyperchromasia, nuclear enlargements, plurinuclearity, nuclear pleomorphism and prominent nucleoli, while cytologic atypia exhibit cellular pleomorphism, elevated cytoplasm volume levels, giant tumor cells and nuclear cytoplasm invaginations.
In order to come to the diagnosis “LMS”, at least two of the following characteristics must be fulfilled:
–an MI of ≥ 10 M/10 HPF (significantly lower in epithelioid and myxoid LMS; see below);
–significant diffuse or multifocal moderate-to-severe atypia.
Where TCN have been identified beyond doubt and there is any degree of mitotic activity, i.e. anyMI, the diagnosis must be LMS, even in the absence of nuclear and cellular atypia. If the typical necroses are absent, then the MI must be ≥ 10 M/10 HPF and, additionally (!), there must also be diffuse or multifocal moderate-to-severe atypia in order to allow the tumor to be classified as LMS. Even when the diagnosis is already clear on the basis of the necroses and atypia, the MI should nonetheless always be noted so as to be able to determine prognosis according to the overall survival probability prediction tool for LMS (329). Particular features regarding MI apply to epithelioid and myxoid LMS (see epithelioid and myxoid LMS). Since MI is 1.5 ± 0.5 higher when mitoses are counted using PHH 3-IHC compared to H&E staining (229), the pathologist needs to record which method was used for counting mitoses. This is vital in order to be able to discern LMS from STUMP, mitotically active LM and bizarre LM, as well as for determining the prognostic score according to Zivanovic (329). An analysis (229) on 132 smooth muscle uterine tumors (26 LMS, 16 STUMP, 30 bizarre LM, 30 mitotically active LM and 30 regular LM) revealed that, with the exception of LM, MI determined via PHH 3-IHC and H&E staining did not really differ from each other.
In rare cases, LMS or their metastases can contain undifferentiated areas with highly pleomorphous cells without signs of smooth muscle differentiation. LMS can also contain sections with a rhabdomyosarcomatous or liposarcomatous differentiation. In such cases, differential diagnostics can be particularly tricky. Up to 70% of LMS also contain foci that are reminiscent of regular LM. While these foci do not have an elevated proliferation index (Ki67 index) and do not express p53, they do contain the genetic characteristics of LMS (198).
The findings pertaining to whether or not LMS express ER and PGR are contradictory.While some have reported that both ER and PGR are expressed either not at all or only very mildly (59), another study has revealed positivity for ER expression in 100% of the LMS in the sample (13 cases), though without making any statement regarding the intensity of said expression. These tumors did not express PGR (133). In a recent study, ER expression was positive in 63% of LMS patients, while the PGR were positive in 50%; 29% of tumors were negative in terms of both ER and PGR expression (122). Both ER and PGR can be verified in 20–60% of cases (Fig. 2.1.3 (E)), though the receptors are usually only focally distributed and predominantly situated near the margins of the tumor (9, 32, 326). In any case, there are substantial differences to LM– ER, PGR and AR are expressed in 78, 88 and 32% of LM, respectively, while the figures for LMS are only 40, 38 and 40%, respectively (171). In LMS, HR expression is independent of disease stage (171). There is an interesting account in which the metastases of an HR negative primary tumor showed a strong immunoreaction for both ER and PGR (59). In summary, a strong decrease or complete loss of PGR expression in particular (but also ER expression) can be regarded as a factor (albeit weakly) suggestive of LMS in the context diagnostically differentiating between LMS and LM.
Immunohistochemical line differentiation needs to be assessed for all mesenchymal tumors. LMS are usually positive for the expression of myogenic markers like desmin and h-caldesmon (Fig. 2.1.3 (F), (G)). In cases of largely undifferentiated LMS, determining the precise tumor type at hand will usually require that all availablemarkers are drawn on.