Soft tissue sarcomas (STSs) represent a heterogeneous group of rare malignant tumors with a wide spectrum in terms of histologic type and prognosis.1 Histologic classification has some limitations and is not very convenient for patient care. It is complex and subject to periodical variations, and its reproducibility is relatively poor among pathologists who are not familiar with soft tissue lesions.2-4 Moreover, about 30% of sarcomas remain unclassified or enter the malignant fibrous histiocytoma group, and except for some rare types, the prognostic value of this histogenetic classification is questionable.
In 1977, Russell et al5 proposed the first coherent and effective system. Their clinicopathologic classification separated patients into 4 stages with a different prognosis and introduced a histologic grading applicable to all adult STSs. This grading was the most important factor of their system.
Almost 30 years later, it is admitted that histologic grade is the most important prognostic factor for adult STS. This has been proved by several multivariate studies6-13 and is clearly stated in the World Health Organization classification.1 It is a key parameter in the currently used clinicopathologic staging,14 and it is one of the important recommendations for the pathologic reporting of STS.15
However, grading has been a controversial topic in STS for many years and is not accepted by some pathologists because it is used on the whole group of STSs considered as a single entity.16,17 Moreover, the value of grading is limited on core needle biopsies, which are being used more and more for diagnosing STS, and new developments in molecular genetics and targeted therapies question the future of classical grading.
DESCRIPTION OF GRADING
In the Russell et al5 study, histologic grade was rather subjectively based on cellularity, cellular pleomorphism, and mitotic activity. To obtain the most efficient system, histologic parameters should be selected by multivariate analysis to use only the necessary parameters summarizing all the prognostic histomorphologic information. According to the studies that have used this methodology, the best parameters are histologic type and subtype and/ or differentiation, tumor necrosis, mitotic index,18-20 and, in a few studies, vascular invasion.21 Some histologic parameters may be very impressive but have never been selected by statistical analysis, and those such as cellularity, nuclear atypia, and pleomorphism should not be considered for grading STS.
Several grading systems have been reported and a review with critical analysis of the most important systems proposed was performed by Oliveira and Nascimento.22 Currently the most common systems used are the French grading18,23 and the National Cancer Institute grading.19,24
Both are 3-grade systems and are quoted in the latest edition of the World Health Organization classification of STS.1 The National Cancer Institute system is based on tumor histologic type and subtype, location, and the amount of tumor necrosis, but cellularity, nuclear pleomorphism, and mitotic index are also considered for some tumor types.
The French grading system is based on 3 parameters (Table 1): tumor differentiation, mitotic index, and tumor necrosis. These parameters are scored 1 to 3 for differentiation and mitotic index and 0 to 2 for necrosis. A 3-grade system is obtained by summing the scores obtained for each of these 3 parameters. Grade 1 is defined as a total of 2 or 3; grade 2 as a total of 4 or 5; and grade 3 as a total of 6 to 8. To enhance the reproducibility of the system, the parameters are defined as exactly as possible. The rules for using this system and its limitations have been previously described in detail.25 Differentiation is the most controversial parameter and is in fact a mixture of histologic type and subtype and/or true differentiation. This parameter was modified in 1997.23 Table 2 shows the differentiation scores attributed to common sarcoma types. A score of 1 is currently assigned to sarcomas closely resembling normal adult tissue to such a degree as to be confused with benign tumors, such as a well-differentiated leiomyosarcoma. A score of 3 is given to embryonal and poorly-differentiated sarcomas, sarcomas of doubtful histologic type, synovial sarcoma, primitive neuroectodermal tumor, osteosarcoma, pleomorphic rhabdomyosarcoma, and pleomorphic liposarcoma. Other sarcomas of certain histologic type such as myxoid liposarcoma are scored 2. Mitotic count is obtained in 10 successive highpower fields (HPFs) in most mitotic areas. This count is taken to establish the score: score 1 for 0 to 9 mitoses; score 2 for 10 to 19 mitoses; and score 3 for more than 19 mitoses per 10 HPF. Strict rules should be followed for counting mitoses: early and adequate fixation of tumor, correct macroscopic sampling (1 section per 1 or 2 cm of the tumor diameter), good quality sections, and adequate selection of the most mitotic areas at medium power and where mitoses should be counted. A calibrated HPF surface should be performed according to the first study (HPF = 0.174 mm2).18 Moreover, pathologists should take their time as there is no use in hurriedly counting mitoses because reproducibility will dwindle. If the mitotic count is close to the cutoff score (ie, 7-9 or 17-19 mitoses per 10 HPF), recounting is strongly advised. Any necrotic or hypocellular areas as well as zones of ulceration should be avoided. Tumor necrosis should be evaluated at histologic and macroscopic levels. However, macroscopic evaluation of necrosis is poorly reproducible and should always be checked under the microscope. Necrosis related to previous surgery or to ulceration must not be taken into account. Hemorrhagic changes or hyalinization should not be considered either. The score is 0 when there is no necrosis, 1 when the necrotic area is less than 50%, and 2 when it is more than 50% of the tumor surface. When the reproducibility of this grading system was tested by 15 pathologists on 25 cases, the crude proportion in agreement was 75% for tumor grade, 73% for mitotic index, 74% for differentiation, and 81% for tumor necrosis.26 Reproducibility of diagnosis of histologic type was significantly lower with an agreement rate of 61%.
PRACTICAL VALUE OF GRADING
Prognostic Value of Grading
Prognosis of STS is dominated by local recurrence and distant metastasis. STSs recur locally in 20% to 30% and metastasize in 30% to 50% of cases.
In most reported studies, quality of surgical margins is the most important factor for predicting local recurrence. 6-10 However, grade has also been reported as a predictor of local recurrence in a few studies.6,27
In fact, histologic grade particularly indicates the probability of distant metastasis and overall survival. Several studies with a multivariate analysis on metastasis have been published.6-13 In every study, whatever the system used, grade is the most important factor. In a study on 1240 patients with a localized STS,12 the metastasis-free 5-year survival rate was 90.8% for grade 1, 71.4% for grade 2, and 43.5% for grade 3 (Figure). In both this study and in others, multivariate analysis for predicting metastatic risk also retained tumor size, neurovascular or bone involvement, and tumor depth in addition to histologic grade. These studies underline the importance not only of grade but also of clinical factors.
The prognostic value of the National Cancer Institute and French grading systems were compared in a series of 410 patients with follow-up.23 In monovariate analysis, both systems were of good prognostic value in predicting overall and metastasis-free survival, with very similar curves. However, distribution of grade according to the 2 systems showed discrepancies in 34.6% of cases, and in the National Cancer Institute system, there were more patients in the noninformative category of grade 2 tumors and fewer grade 3 tumors. In multivariate analysis, when the systems were compared in the same model, French grade 3 ranked first in both overall and metastasis-free survival situations.
Contribution of Grading to Patient Management
Because the main value of grading is to predict the advent of metastasis, its main value for therapeutic management is the most efficient use of chemotherapy. The response rate to palliative chemotherapy is better in patients with high-grade sarcoma than in those with lower grade tumors.28 Adjuvant chemotherapy has also been reported to be more efficient in grade 3 tumors than in grade 1 or 2 tumors.27 In 3-grade systems, histologic grade is basically helpful in the management of patients with grade 1 and grade 3 only. Patients with a grade 1 tumor should not receive chemotherapy because it is useless and inefficient. On the other hand, chemotherapy could be useful and efficient in patients with a grade 3 tumor. The grade 2 category, which accounts for 40% to 50% of cases, is noninformative for prognosis and therapeutic decision and comprises a mixture of tumors of good and bad prognosis. For this group of patients, knowledge of clinical factors is important. Ravaud et al29 showed that the prognosis of superficial grade 2 tumors is almost the same as that of grade 1 tumors. Conversely, deep-seated grade 2 tumors and grade 3 tumors were high-risk lesions, for which preoperative or adjuvant chemotherapy would be indicated. Another study27 showed that among patients with a grade 2 tumor, adjuvant chemotherapy was more efficient in patients with a deep tumor larger than 5 cm as compared to those with a superficial tumor or one smaller than 5 cm.
LIMITATIONS OF GRADING AND CONTROVERSIAL POINTS
Conditions of Use of Grading
Grading does not differentiate benign and malignant lesions, and before grading a soft tissue lesion, one must be sure that one is dealing with a true sarcoma and not a pseudosarcomatous lesion or a nonmesenchymal tumor such as a carcinoma, a melanoma, or a lymphoma. Sarcoma-like lesions, such as nodular and proliferative fasciitis, display some characteristics of malignancy, and these lesions could be of high grade if grading were used. STS grading is not applicable to all kinds of sarcomas. Grading has been shown to be a predictive factor in STS, but its prognostic value in sarcomas of other locations, such as the female genital tract and gastrointestinal tract, is not proven. Grading should be performed on untreated tumors, because radiotherapy and chemotherapy can increase tumor necrosis and decrease mitotic activity.
Grading and Typing
The respective value of grading and histologic type is the most controversial point.16,17,30,31 Grading is applied to the sarcoma group as a whole, but some pathologists believe that it cannot be applied to the different types of sarcomas because its predictive value differs according to histologic type. They therefore believe that a grading system should be adapted to each histologic type, making grading unfeasible in daily practice given the number of histologic types in adult STS.
In a previous study,12 we looked at the prognostic value of grading in comparison with the clinical parameters in each of the main histologic types in a series of 1240 nonmetastatic STSs. There were no significant differences between any of the grade 1 tumors, any of the grade 2 tumors, and any of the grade 3 tumors in each of the main histologic types. Multivariate analysis performed in each of these histologic types showed that grade was the most important factor for predicting metastatic risk in so-called malignant fibrous histiocytoma, unclassified sarcomas, synovial sarcomas, and miscellaneous sarcomas considered as a single group. It is the second most important factor in leiomyosarcomas after bone or neurovascular invasion and the third most important in liposarcomas after tumor size and histologic subtype. Grade was of no prognostic value in malignant peripheral nerve sheath tumors and rhabdomyosarcomas, but no factor could be selected in multivariate analysis in these 2 small groups of patients. In rhabdomyosarcomas, monovariate analysis showed that grade had a significant impact on metastasis-free survival.
Grading is also less informative than histologic type in dedifferentiated liposarcomas, round cell liposarcomas, and alveolar soft part sarcomas. Of course, it should not be used on tumors of intermediate malignancy such as atypical fibroxanthoma and dermatofibrosarcoma protuberans. Grading is not at all suitable for atypical fibroxanthoma, because it would paradoxically be a high-grade lesion, even though it is a virtually nonmetastatic tumor. Despite these restrictions, tumors for which histologic grade is poorly informative represent only a small part of sarcomas, certainly less than 10% of all adult STSs.
In that study,12 univariate analysis showed that grade and histologic type were predictive factors for metastasis outcome. However, multivariate analysis retained grade but not histologic type, which had no independent prognostic value. In fact, grade partly depends on histologic type with differentiation score, and therefore, the prognostic value of histologic type is already included in the grade.
Grading in Childhood STS
The most important prognostic factor in pediatric STS is histologic type: is the tumor a rhabdomyosarcoma, a primitive neuroectodermal tumor, or not? Among rhabdomyosarcomas, histologic subtype is considered as the best prognostic factor, with a good prognosis for the botryoid and the spindle cell subtypes and a poor prognosis for the alveolar subtype.32 Primitive neuroectodermal tumor is considered as a high-grade tumor. Grading has been used with success in other childhood sarcomas. The Pediatric Oncology Group grading system established by Parham et al33 is based on outlines proposed by Costa et al19 and Enzinger and Weiss34 and contains 3 main parameters: histologic type, tumor necrosis, and mitotic activity. Specific histologic subtypes with a low tendency for malignant behavior were graded 1, any sarcoma not automatically designated as grade 1 or 3 with tumor necrosis less than 15% and mitotic activity of 5 or less mitoses per 10 HPF were graded 2, and specific histologic subtypes with known aggressive behavior and/or any sarcoma with tumor necrosis of more than 15% or mitotic activity of more than 5 mitoses per 10 HPF were graded 3. In this system, the age of the patient is also taken into account. This grading system has a good prognostic value35,36 with almost 100% for metastasis-free survival for grades 1 and 2 and only 50% for grade 3 in 1 study.36
Grading Using Fine-Needle Aspiration and Core Needle Biopsies
Grading is recommended on representative material, that is, on the whole tumor or on an incisional biopsy. However, core needle biopsies or even fine-needle aspirations are being more widely used for the primary diagnosis of an STS, with an increasing use of preoperative chemotherapy. Core needle biopsies are more suitable than fine-needle aspirations for the diagnosis of sarcoma, histologic typing, and grading.37-39 Several previous studies reported a good concordance of grading between core needle biopsies and incisional biopsies.40-42 However, it is difficult to assess necrosis and mitotic index on this material, and the grading system used in these studies was not specified. Moreover, core needle biopsies are prone to sampling error with the risk of underestimation of grade, and unless obvious high-grade tumor is present, it is often inappropriate to use classical grading systems on core needle biopsies.43
The Future of Grading
Grading should certainly be adapted to the modern management of patients and should be complemented with radiologic and molecular parameters.
More and more patients receive preoperative treatment such as chemotherapy, radiotherapy, or tumor necrosis factor after a microbiopsy. Therefore, grading should be adapted to this type of material. Extension of necrosis could be evaluated by imaging studies, which can be viewed as a type of macroscopic examination. Computed tomography, magnetic resonance imaging,44,45 and other techniques such as Doppler ultrasonography with perfusion software and contrast agent injection,46 31Phosphorus-magnetic resonance spectroscopy,47 and positron emission tomography using 2-fluoro-2-deoxy-D-glucose48 have been reported to be able to assess tumor necrosis, chemotherapy response, and grading in STS. Mitotic index, which is difficult to evaluate on core needle biopsies, could be replaced by a mib1 score, as was recently the case when Hasegawa et al used the French grading system. These authors showed that grading using the mib1 score is more reproducible49 and has a better predictive value than the classical grading using mitotic index.50 The prognostic value of the MIB1 score should certainly be tested on core needle biopsies.
Grading can be considered as a morphologic translation of molecular events that determine tumor aggressiveness. Therefore, molecular parameters could eventually complement or even replace histologic parameters. Ladanyi et al51 reported the impact of the transcript fusion type on the prognosis of synovial sarcomas and reported a significantly better prognosis for patients with a SSX2 synovial sarcoma as compared to patients with an SSX1 one. A subsequent study52 did not confirm these results and further studies are necessary. In alveolar rhabdomyosarcomas, the fusion transcript type has been reported as having a prognostic value with a better prognosis for patients with a PAX7 tumor as compared with patients with a PAX3 tumor.53,54 In gastrointestinal stromal tumors, the type of mutation has been reported as the best predictor for imatinib response.55,56
CONCLUSIONS
Histologic grade is currently the most important prognostic factor and the best indicator of metastatic risk for adult STS. Therefore, it should be part of the pathology report with histologic type and subtype, tumor size, tumor depth, and surgical margins. It should be adapted to the modern management of patients and should be complemented with radiologic parameters when dealing with a microbiopsy. Given the recent strides forward in molecular genetics and the concept of targeted therapy, grading could eventually be complemented or even replaced by molecular parameters.
© 2006 College of American Pathologists Provided by ProQuest LLC. All Rights Reserved.
