Old versus new FIGO staging systems in predicting overall survival in patients with uterine leiomyosarcoma: a study of 86 cases

Development of a Multi-Institutional Prediction Model for Three-Year Survival Status in Patients with Uterine Leiomyosarcoma (AGOG11-022/QCGC1302 Study)

Simple Summary

Uterine leiomyosarcoma is an aggressive tumor and the current staging system cannot differentiate the patients into different prognostic groups. This leads to difficulty in predicting the patients’ outcomes and planning for adjuvant therapy. We aimed to develop a prediction model that can predict the chance of survival by the third year. In this article, we had used different statistical tests to identify five readily available clinicopathologic parameters to build the prediction model. Internal validation was performed with satisfactory accuracy. Such a prediction model might help to predict survival outcome, and guide future research on the treatment modality.

Abstract

Background: The existing staging systems of uterine leiomyosarcoma (uLMS) cannot classify the patients into four non-overlapping prognostic groups. This study aimed to develop a prediction model to predict the three-year survival status of uLMS. Methods: In total, 201 patients with uLMS who had been treated between June 1993 and January 2014, were analyzed. Potential prognostic indicators were identified by univariate models followed by multivariate analyses. Prediction models were constructed by binomial regression with 3-year survival status as a binary outcome, and the final model was validated by internal cross-validation. Results: Nine potential parameters, including age, log tumor diameter, log mitotic count, cervical involvement, parametrial involvement, lymph node metastasis, distant metastasis, tumor circumscription and lymphovascular space invasion were identified. 110 patients had complete data to build the prediction models. Age, log tumor diameter, log mitotic count, distant metastasis, and circumscription were significantly correlated with the 3-year survival status. The final model with the lowest Akaike’s Information Criterion (117.56) was chosen and the cross validation estimated prediction accuracy was 0.745. Conclusion: We developed a prediction model for uLMS based on five readily available clinicopathologic parameters. This might provide a personalized prediction of the 3-year survival status and guide the use of adjuvant therapy, a cancer surveillance program, and future studies.

Copper ions are novel therapeutic agents for uterine leiomyosarcoma

BACKGROUND

Multidrug resistance is a major concern in uterine leiomyosarcoma treatment. Development of effective chemotherapies and management of drug resistance in patients is necessary. The copper efflux transporter adenosine triphosphatase copper transporting beta is a member of the P-type adenosine triphosphatase family and is also known as a strong platinum efflux transporter. Various reports have shown the association between adenosine triphosphatase copper transporting beta and platinum resistance; however, suitable inhibitors or methods for inhibiting platinum efflux via adenosine triphosphatase copper transporting beta are not developed.

OBJECTIVE

Our study focused on platinum resistance in uterine leiomyosarcoma. The role of adenosine triphosphatase copper transporting beta in uterine leiomyosarcoma resistance to platinum drugs was investigated both in vitro and in vivo.

STUDY DESIGN

Adenosine triphosphatase copper transporting beta expression was investigated by Western blotting and the efficacy of copper sulfate pretreatment and cisplatin administration in adenosine triphosphatase copper transporting beta-expressing cells was investigated both in vitro and in vivo.

RESULTS

Western blot analysis of SK-LMS-1 cells (uterine leiomyosarcoma cell line) revealed strong adenosine triphosphatase copper transporting beta expression. A permanent SK-LMS-ATPase copper transporting beta-suppressed cell line (SK-LMS-7B cells) was generated, and cisplatin exhibited a significant antitumor effect in SK-LMS-7B cells, both in vitro (SK-LMS-1 cells, half-maximal inhibitory concentration, 17.2 μM; SK-LMS-7B cells, half-maximal inhibitory concentration, 4.2 μM, P < .01) and in xenografts compared with that in SK-LMS-1 cells (5.8% vs 62.8%, P < .01). Copper sulfate was identified as a preferential inhibitor of platinum efflux via adenosine triphosphatase copper transporting beta. In SK-LMS-1 cells pretreated with 15 μM copper sulfate for 3 hours, the cisplatin half-maximal inhibitory concentration decreased significantly compared with that in untreated cells and resulted in significantly increased intracellular platinum accumulation (1.9 pg/cell vs 8.6 pg/cell, P < .01). The combination of copper sulfate pretreatment with cisplatin administration was also effective in vivo and caused cisplatin to exhibit significantly increased antitumor effects in mice with SK-LMS-1 xenografts (3.1% vs 62.7%, P < .01).

CONCLUSION

Our study demonstrates that adenosine triphosphatase copper transporting beta is overexpressed in uterine leiomyosarcoma cells and that copper sulfate, which acts as an inhibitor of platinum efflux via adenosine triphosphatase copper transporting beta, may be a therapeutic agent in the treatment of uterine leiomyosarcoma.

The combination of histone deacetylase and p53 expressions and histological subtype has prognostic implication in uterine leiomyosarcoma

OBJECTIVE

To investigate the expression of different histone deacetylases and their association with disease characteristics and survival outcomes in uterine leiomyosarcoma patients.

METHODS

The immunohistochemical expression of different histone deacetylases and p53 by tissue microarray and histological subtypes were assessed in tumor tissue samples of 42 eligible patients.

RESULTS

Histone deacetylases 1-4, 6 and 8 showed prevalent and strong (3+) expression (88.1, 90.5, 95.2, 92.9, 83.3 and 100%, respectively). Histone deacetylases 5, 7 and 9 showed infrequent strong expression (33.3, 50 and 38.1%, respectively). There were trends of higher disease-free survival rates according to the combination of weaker expression of histone deacetylase 5, 7 or 9 with positive p53 expression or with non-epithelial subtype. The patients with triple-positive favorable prognostic factors (any of weaker histone deacetylase 5, 7 and 9 expression, p53 positive, and non-epithelioid subtype) had the better survival outcomes while the patients with other combinations had the worse survival outcomes. In multivariate analysis, histone deacetylase 5 in combination with epithelioid subtype was an independent predictor for disease-free survival.

CONCLUSIONS

Expression of histone deacetylase 5, 7 and 9 is a potential prognostic marker in uterine leiomyosarcoma when combined with pathologically relevant prognostic factors (p53 and histological subtype). This prevalent and strong histone deacetylase expression warrants further study in well-designed investigations of histone deacetylases as therapeutic targets in uterine leiomyosarcoma.

Androgen receptor as a prognostic biomarker and therapeutic target in uterine leiomyosarcoma

OBJECTIVE

To investigate the expression of androgen receptor (AR) and its correlation with disease status and survival outcome in uterine leiomyosarcoma with other hormone receptors.

METHODS

The medical records and paraffin blocks of 42 patients were reviewed. The immunohistochemical expression of AR, estrogen receptor (ER), progesterone receptor (PR), gonadotropin releasing hormone (GnRH), and cytochrome P450, family 19, subfamily A, polypeptide 1 (CYP19A1) were assessed using tissue microarray.

RESULTS

In total, AR expression was observed in 11 patients (26.2%). International Federation of Gynecology and Obstetrics (FIGO) stage and AR were independent factors for disease-free survival (DFS) in multivariate regression analysis (odds ratio [OR]=5.8; 95% confidence interval [CI]=1.2-28.4 and OR=0.2; 95% CI=0.05-0.90; p=0.029 and 0.032, respectively). There were no deaths in the AR expression group, whereas the 5-year overall survival (OS) was 54.8% in the no expression group (p=0.014). Co-expression of ER and/or PR with AR was associated with significantly better 5-year DFS and OS than those with negative AR (72.7% vs. 28.6% and 100% vs. 64.3%; p=0.020 and 0.036, respectively). AR may be an independent prognostic marker regardless of ER/PR.

CONCLUSION

AR can be a potential prognostic biomarker in uterine leiomyosarcoma.

Practical issues in uterine pathology from banal to bewildering: the remarkable spectrum of smooth muscle neoplasia

Among mesenchymal tumors of the uterus, smooth muscle neoplasms are most common. The wide morphologic spectrum, especially within the category of leiomyomas, is responsible for diagnostic problems more frequently with leiomyosarcoma (including mitotically active, apoplectic, and leiomyoma with bizarre nuclei) but also with endometrial stromal tumors. In the former scenario, clinical information, gross appearance as well as strict utilization of morphologic criteria including cytologic atypia, mitotic activity, and tumor cell necrosis are clues in establishing the correct diagnosis. It is important to keep in mind that mitotic rate thresholds vary for the different subtypes of leiomyosarcoma. Of note, p16 should be used with caution in supporting a diagnosis of leiomyosarcoma as it is often positive in leiomyomas with bizarre nuclei and leiomyomas with apoplectic change (in the latter most frequently and more intense near areas of necrosis). MED12 mutations have also a very limited role in this differential diagnosis. Endometrial stromal tumors are by far, less common than smooth muscle tumors, but can be confused with leiomyosarcomas if they are associated with an undifferentiated uterine sarcoma and the low-grade component is overlooked or they have a myxoid/fibroblastic morphology. The differential diagnosis may be confounded if the latter is associated with a high-grade endometrial stromal sarcoma. It is important to highlight that CD10 is not a reliable marker in these differentials and should be used as a part of a panel of antibodies that also includes desmin and h-caldesmon. Two other recently categorized tumors in the uterus that merit special mention are PEComa and inflammatory myofibroblastic tumor as they enter in the differential diagnosis of smooth muscle tumors. PEComa may be part of the tuberous sclerosis syndrome and may show either a predominantly epithelioid or spindle morphology or combination thereof. Rarely, it may contain melanin pigment. There is variable positivity for HMB-45, MelanA, MiTF, and CathepsinK, and some tumors have been shown to express TFE-3 especially when associated with "clear cell" morphology. Patients with adverse outcome have tumors with ≥2 of the following features: ≥5 cm, infiltration, high-grade cytologic features, mitotic rate ≥1/50 high-power fields, necrosis, or lymphovascular invasion. Inflammatory myofibroblastic tumor is important to recognize as it often mimics myxoid smooth muscle tumors, either benign or malignant. The presence of an associated lymphoplasmacytic infiltrate should alert to that possibility and ALK studies (immunostain or FISH) are helpful in establishing this diagnosis. These tumors can behave in a malignant manner if large, associated with abundant myxoid change, brisk mitotic rate or show tumor cell necrosis.

Cost-Effectiveness of Laparoscopic Hysterectomy With Morcellation Compared With Abdominal Hysterectomy for Presumed Myomas

STUDY OBJECTIVE

Hysterectomy for presumed leiomyomata is 1 of the most common surgical procedures performed in nonpregnant women in the United States. Laparoscopic hysterectomy (LH) with morcellation is an appealing alternative to abdominal hysterectomy (AH) but may result in dissemination of malignant cells and worse outcomes in the setting of an occult leiomyosarcoma (LMS). We sought to evaluate the cost-effectiveness of LH versus AH.

DESIGN

Decision-analytic model of 100 000 women in the United States assessing the incremental cost-effectiveness ratio (ICER) in dollars per quality-adjusted life-year (QALY) gained (Canadian Task Force classification III).

SETTING

U.S. hospitals.

PATIENTS

Adult premenopausal women undergoing LH or AH for presumed benign leiomyomata.

INTERVENTIONS

We developed a decision-analytic model from a provider perspective across 5 years, comparing the cost-effectiveness of LH to AH in terms of dollar (2014 US dollars) per QALY gained. The model included average total direct medical costs and utilities associated with the procedures, complications, and clinical outcomes. Baseline estimates and ranges for cost and probability data were drawn from the existing literature.

MEASUREMENTS AND MAIN RESULTS

Estimated overall deaths were lower in LH versus AH (98 vs 103). Death due to LMS was more common in LH versus AH (86 vs 71). Base-case assumptions estimated that average per person costs were lower in LH versus AH, with a savings of $2193 ($24 181 vs $26 374). Over 5 years, women in the LH group experienced 4.99 QALY versus women in the AH group with 4.91 QALY (incremental gain of .085 QALYs). LH dominated AH in base-case estimates: LH was both less expensive and yielded greater QALY gains. The ICER was sensitive to operative costs for LH and AH. Varying operative costs of AH yielded an ICER of $87 651/QALY gained (minimum) to AH being dominated (maximum). Probabilistic sensitivity analyses, in which all input parameters and costs were varied simultaneously, demonstrated a relatively robust model. The AH approach was dominated 68.9% of the time; 17.4% of simulations fell above the willingness-to-pay threshold of $50 000/QALY gained.

CONCLUSION

When considering total direct hospital costs, complications, and morbidity, LH was less costly and yielded more QALYs gained versus AH. Driven by the rarity of occult LMS and the reduced incidence of intra- and postoperative complications, LH with morcellation may be a more cost-effective and less invasive alternative to AH and should remain an option for women needing hysterectomy for leiomyomata.

Biomarker resolution of uterine smooth muscle tumor necrosis as benign vs malignant

Uterine leiomyosarcomas are rare malignant tumors with a poor prognosis while leiomyomas are common benign tumors unrelated to their malignant counterparts. Diagnostic features commonly present in leiomyosarcoma include cytologic atypia, high mitotic index, and a sarcoma-specific geographic cell death designated 'tumor cell necrosis (TCN)'. TCN has a sharp viable-nonviable boundary lacking inflammation, fibrosis or granulation tissue seen in nonspecific infarction. These characteristics are sometimes difficult to interpret on routine hematoxylin and eosin slides, and can lead to diagnostic errors. In this study, we used extracellular matrix stains to test the hypothesis that the host response which characterizes nonspecific infarction may degrade the matrix in infarcted tumor more than in TCN. A 'honeycomb' pattern of reticulin highlighted individual tumor cells in viable regions of all cases. Nonviable area of reticulin patterns differed significantly by diagnosis (P<0.001), with a honeycomb pattern maintained (91%, 20/22) in leiomyosarcoma and lost (61%, 11/18) in leiomyomas. Retention of honeycomb reticulin in nonviable areas of leiomyosarcoma occurred irrespective of the presence of inflammation, hemorrhage, fibrosis, or diffuse hyalinization. Fibrosis/hyalinization as evidenced by trichrome stain was significantly (P<0.001) more common in nonviable areas of benign leiomyomas (100%, 18/18) compared with leiomyosarcomas (36%, 8/22). In those occasions where viable tissues contained discernable polarization of mitotic activity, these decreased toward the nonviable interface in leiomyosarcoma, and had an opposite pattern in leiomyomas, increasing toward the interface. There is a significant difference in the reticulin and collagen networks of nonviable areas of leiomyosarcoma compared with leiomyoma. At the time of early injury, both retain reticulin; however, this is cleared over time in benign, but not malignant, areas of necrosis. We conclude that proliferative repair of leiomyomas at the viable-nonviable interface includes remodeling of the extracellular matrix, in contrast to the static preservation of extracellular matrix ('mummification') in nonviable areas of leiomyosarcomas.

Prognostic factors and survival in patients treated surgically for recurrent metastatic uterine leiomyosarcoma

Background. Uterine leiomyosarcoma (LMS) is a rare diagnosis, which is seldom cured when it recurs with metastatic disease. We evaluated patients who present with first time recurrence treated surgically to determine prognostic factors associated with long-term survival. Methods. Over a 16-year period, 41 patients were operated on for recurrent uterine sarcoma. Data examined included patient age, date of initial diagnosis, tumor histology, grade at the initial diagnosis, cytopathology changes in tumor activity from the initial diagnosis, residual tumor after all operations, use of adjuvant therapy, dates and sites of all recurrences, and disease status at last followup. Results. 24 patients were operated for first recurrence of metastatic uterine LMS. Complete tumor resection with histologic negative margins was achieved in 16 (67%) patients. Overall survival was significantly affected by the FIGO stage at the time of the initial diagnosis, the ability to obtain complete tumor resection at the time of surgery for first time recurrent disease, single tumor recurrence, and recurrence greater than 12 months from the time of the initial diagnosis. Median disease-free survival was 14 months and overall survival was 27 months. Conclusion. Our findings suggest that stage 1 at the time of initial diagnosis, recurrence greater than 12 months, isolated tumor recurrence, and the ability to remove ability to perform complete tumor resection at the time of the first recurrence can afford improved survival in selected patientsat the time of the first recurrence can afford improved survival in selected patients.

Recent developments in uterine mesenchymal neoplasms

Smooth muscle and endometrial stromal tumours represent the two most common uterine mesenchymal neoplasms that may present diagnostic dilemmas for the practising surgical pathologist. Recent changes in morphological and staging criteria, as well as the discovery of new immunohistochemical markers, have improved the diagnosis and classification of these tumours. We highlight the difficulty in distinguishing tumour cell necrosis from infarct-type necrosis and the limited utility of p16 immunohistochemical expression in the diagnosis of leiomyosarcoma. We also discuss the controversial use of mitotic activity and necrosis as prognostic factors in endometrial stromal sarcomas. Emerging genetic information has also greatly expanded our understanding of 'sarcomagenesis' in both tumour types and may provide insight into potential therapeutic targets for the treatment of leiomyosarcoma and endometrial stromal sarcomas, harboring MED12 (mediator complex subunit 12) mutations and recurrent gene rearrangements, respectively. In this review, we discuss the core updates in the diagnosis and classification of uterine leiomyosarcomas and endometrial stromal sarcomas, highlighting new and important molecular genetic findings that may drive pathogenesis.