THE COMPARATIVE ROLES OF
ELECTRON MICROSCOPY AND IMMUNOHISTOCHEMISTRY
IN THE DIAGNOSIS OF POORLY DIFFERENTIATED SARCOMAS

Object of the Study
The histologic typing of soft tissue sarcomas is generally feasible when dealing with well or moderately differentiated tumors, such as biphasic synovial sarcoma, fibrosarcoma, liposarcoma or extraskeletal bone and cartilage tumors, which are characterized by a typi-cal morphologic profile. The basic immunohistochemical staining patterns of the different types of well/moderately differentiated sarcomas are similarly well-established. Con-versely, poorly differentiated soft tissue sarcomas quite often represent a difficult diagnostic task, in large part due to the lack of specificity of diagnostic histologic and immunohis-tochemical criteria. The aim of our study was to investigate the comparative roles of elec-tron microscopy and immunohistochemistry in the proper identification of poorly differenti-ated, spindle cell and pleomorphic soft tissue sarcomas.

Materials & Methods
We investigated 34 cases of poorly differentiated, spindle cell and pleomorphic sar-comas arising in deep soft tissues (excluding the retroperitoneum) of adult patients ob-served in our Institution in the last five years. This series represents the 13.4% of the total amount of soft tissue sarcomas observed in the same period of time (253 cases). Eighteen cases were investigated by means of immunohistochemistry and electron microscopy from the beginning while, in 16 cases, the patients—who had already underwent a biopsy or a surgical procedure in another institution—were hospitalized to excise the tumor or to cure a recurrence. In these cases, we were initially able to study only hematoxylin & eosin stained slides, due to the unavailability of material for immunohistochemical or submicro-scopic investigations. The diagnosis made on hematoxylin & eosin stained sections was successively compared with that formulated following the more exhaustive study per-formed on the excised specimens by means of immunohistology and electron microscopy.

Results
The 18 cases exhaustively investi-gated from the beginning consisted of: 3 fibrosarcomas, 4 leiomyosarcomas, 2 rhab-domyosarcomas, 3 liposarcomas, 4 monophasic synovial sarcomas, 1 malignant pe-ripheral nerve sheet tumor, and 1 myoftbrosarcoma. The final diagnoses—along with the comparison with the original ones—in the other 16 cases are reported in Table I.

Table I. Comparison between original diagnosis (lines) and final diagnosis (columns)

	FS	MFS	LMS	LPS	SS
FS	-	-	-	-	1
MFH	2	3	4	3	-
LPS	-	2	-	-	-
MPNST	-	1	-	-	-

FS: fibrosarcoma, MFH: malignant fibrous histiocytoma, MFS: myofibrosarcoma, LMS: leiomyosarcoma, LPS: liposarcoma, MPNST: malignant peripheral nerve sheath tumor, SS: monophasic synovial sarcoma.

In no case, after immunohistochemistry and electron microscopy, was the original diag-nosis confirmed, especially for malignant fibrous histiocytoma.

The more salient submicroscopic features of the tumors investigated are synopti-cally summarized in Table II, and the immunohistochemical data in Table III.

Table II. Salient submicroscopic features (%) according to the histotype

	FS (N=5)		MFS (N=7)		LMS (N=8)		RMS (N=2)		LPS (N=6)		MPNST (N=1)		SS (N=5)
Feature	+	++	+	++	+	++	+	++	+	++	+	++	+	++
RER	60.0	40.0	57.1	42.9	37.5	-	50.0	50.0	33.3	66.7	100	-	40.0	-
Lipid vacuoles	40.0	-	28.6	-	12.5	-	50.0	-	-	100	-	-	-	-
Glycogen	-	-	14.3	-	12.5	-	-	-	-	16.7	-	-	-	-
Intermediate filaments	40.0	-	71.4	-	50.0	-	100	-	33.3	-	100	-	40.0	40.0
Thin filaments	-	-	42.9	57.1	25.0	75.0	-	50.0	-	-	-	-	-	-
Sarcomeres	-	-	-	-	-	-	50.0	50.0	-	-	-	-	-	-
Micropinocytotic vesicles	20.0	-	71.4	-	25.0	37.5	100	-	-	-	-	-	20.0	-
Cell junctions	60.0	-	57.1	28.6	62.5	12.5	50.0	-	33.3	-	-	-	20.0	80.0
Basal lamina	20.0	-	71.4	-	37.5	37.5	50.0	-	83.3	16.7	-	-	60.0	-
Fibronexa	-	-	14.3	-	-	-	-	-	-	-	-	-	-	-
Luse bodies	-	-	-	-	12.5	-	-	-	-	-	-	-	-	-

FS: fibrosarcoma, MFS: myofibrosarcoma, LMS: leiomyosarcoma, RMS, rhabdomyosarcoma, LPS: liposarcoma, MPNST: malignant peripheral nerve sheath tumor, SS: monophasic synovial sarcoma. +: Inconspicuous feature, ++: Prominent feature

Table III. Results of the immunohistochemical study (%) according to the histotype

	FS (N=5)		MFS (N=7)		LMS (N=8)		RMS (N=2)		LPS (N=6)		MPNST (N=1)		SS (N=5)
Feature	+	++	+	++	+	++	+	++	+	++	+	++	+	++
Vimentin	20.0	80.0	28.6	71.4	12.5	87.5	-	100	50.0	50.0	-	100	20.0	80.0
S100-protein	-	-	14.3	-	-	-	-	-	16.7	-	-	100	40.0	-
Muscle specific actin	-	-	57.1	28.6	37.5	62.5	50.0	50.0	-	-	-	-	-	-
alpha-smooth muscle actin	-	-	-	-	25.0	62.5	100	-	-	-	-	-	-	-
Desmin	-	-	-	-	12.5	-	50.0	-	-	-	-	-	-	-
Myoglobin	-	-	-	-	-	-	-	-	-	-	-	-	-	-
CD68	20.0	-	-	-	-	-	-	-	50.0	16.7	-	-	-	-
CD34	-	-	14.3	-	-	-	-	-	-	-	-	-	-	-
Collagen IV	-	-	14.3	14.3	12.5	25.0	-	-	16.7	-	-	-	20.0	-
Laminin	-	-	-	-	-	-	-	-	16.7	-	-	-	-	-
Fibronectin	-	-	28.6	28.6	12.5	25.0	-	-	16.7	16.7	-	-	20.0	-
Cytokeratins	-	-	-	-	-	-	-	-	-	-	-	-	40.0	-
EMA	-	-	-	-	12.5	-	-	-	-	-	-	-	40.0	40.0

FS: fibrosarcoma, MFS: myofibrosarcoma, LMS: leiomyosarcoma, RMS, rhabdomyosarcoma, LPS: liposarcoma, MPNST: malignant peripheral nerve sheath tumor, SS: monophasic synovial sarcoma. +: Focal positivity (<20% of tumor cells, ++: Diffuse positivity (>20% of tumor cells)

Comment
Our study clearly documented that ancillary techniques are mandatory in order to properly identify poorly differentiated, spindle cell and pleomorphic soft tissue sarcomas. In fact, in none of the cases seen in consultation, in which we were initially unable to perform electron microscopy and/or immunohistochemistry, the diagnosis originally made—as a rule, in the absence of definite criteria against the proffered diagnosis usually confirming the diagnosis of the referring pathologist—was confirmed upon the more accurate investi-gation using electron microscopy and immunohistochemistry. Notably, malignant fibrous histiocytoma, that was the most represented entity in this series (12 out of 16 cases), does not withstand the categorical scrutiny of electron microscopy and immunohistology, and those cases were reclassified by such means as poorly differentiated examples of fibro-sarcoma, myofibrosarcoma, Ieiomyosarcoma, and liposarcoma.

In our experience, the immunohistochemical analysis does not quite often give clear-cut information concerning the tumor cell phenotype, when dealing with poorly differ-entiated spindle cell and pleomorphic soft tissue sarcomas, and therefore is less valuable than electron microscopy in typing these tumors, as opposed to what generally considered on the basis of the everyday practice when more often dealing with well and moderately differentiated soft tissue sarcomas.

In particular, in our series cytokeratins were documented to be specific for synovial sarcoma, although the sensitivity was low. Conversely, epithelial membrane antigen (EMA) was more sensitive but, unfortunately, not completely specific. Markers identifying contrac-tile filament peptides were restricted to specific tumor categories—specifically, myofibro-sarcoma, Ieiomyosarcoma, and rhabdomyosarcoma—but did not allow a discrimination among them even though, in our experience, a-smooth muscle actin was never docu-mented in myofibrosarcoma. S-ICC protein was present in several tumor types, even though was diffusely expressed only in the only malignant peripheral nerve sheet tumor case. Similarly, CD68, despite not specific, was consistently expressed by Iiposarcomas only in our series. Laminin, fibronectin, and collagen type IV were inconsistently seen, with a predilection for myofibrosarcoma, Ieiomyosarcoma, liposarcoma, and synovial sarcoma.

Concerning electron microscopy, only the presence of rudimentary sarcomeres was found to be specific and sensitive for rhabdomyosarcoma; while the presence of fibronexa, despite considered specific of myofibrosarcoma, was documented only in one such a case. Moreover, electron microscopy was most valuable in sampling out liposarcoma, due to the identification of multiple cytoplasmic lipid vacuoles in a significant number of tumor cells. In addition, the combined evaluation of different submicroscopic features in the sin-gle case (R.E.R., intermediate filaments, thin filaments, micropinocytotic vesicles, cell junctions, basal lamina, fibronexa) allowed the differential diagnosis between fibrosar-coma, myofibrosarcoma, Ieiomyosarcoma, malignant peripheral nerve sheet tumor, and monophasic synovial sarcoma.

In conclusion, the combined use of electron microscopy and immunohistochemistry can allow a proper categorization of poorly differentiated, spindle cell and pleomorphic soft tissue sarcomas in virtually all cases.

References

1. Erlandson RA. Diagnostic Transmission Electron Microscopy of Tumors with Clini-copathological, lmmunohistochemical, and Cytogenetic Correlations. Raven Press, 1994.
2. Eyden BP, Christensen L. Leiomyosarcoma versus myofibrosarcoma: Observations and terminology. Ultrastruot Pathol 1993: 17; 231-239.
3. Fisher C. The value of electron microscopy and immunohistochemistry in the diagnosis of soft tissue sarcomas: A study of 200 cases. Histopathology 1990: 16; 441-454.
4. Fletcher CDM. Pleomorphic malignant fibrous histiocytoma: Fact or fiction? A critical reappraisal based on 159 tumors diagnosed as pleomorphic sarcoma. Am J Surg Pathol 1992:16; 21 3-228.
5. Leong A S-Y, Wick MR, Swanson PE. lmmunohistology and electron microscopy of anaplastic and pleomorphic tumors. Cambridge University Press, I 997.
6. Schurch W, Begin LR, Seemayer TA, et al. Pleomorphic soft tissue myogenic sarco-mas of adulthood: A reappraisal in the mid-1990s. Am J Surg Pathol 1996: 20: 131-
   147.
7. Wick MR, Swanson PE, Manivel JC. lmmunohistochemical analysis of soft tissue sar-comas: Comparison with electron microscopy. Appl Pathol 1988: 6; 169-196.