Context: Central nervous system lesions are diverse and remain one of the most challenging domains for neuropathologists. Intraoperative cytological diagnosis is now a universally accepted technique in diagnosis of central nervous system (CNS) lesions. Aims: 1) To analyze and compare cytomorphological features of CNS lesions in intraoperative squash smears with histopathology, immunohistochemistry, and preoperative radiological diagnosis and 2) to determine the diagnostic accuracy, sensitivity, and specificity of intraoperative squash cytology. Settings and Design: Prospective study conducted at a tertiary healthcare centre over a period of two years. Methods and Material: All biopsy materials which underwent squash cytology and histopathological examination were collected, evaluated, classified, and graded according to WHO classification of CNS Tumors, 2016. The squash cytosmear diagnosis was compared with histopathological features and radiological diagnosis. Discordances were evaluated. Statistical Analysis Used: The cases were categorized into true positives, false positives, true negatives, and false negatives. Diagnostic accuracy, sensitivity, and specificity were calculated from 2*2 table. Results: A total of 190 cases were included in the study. A total of 182 cases (95.70%) were neoplastic of which 87.36% were primary CNS neoplasms. Diagnostic accuracy in non-neoplastic lesions was 88.8%. Most common neoplastic lesions were glial tumors (35.7%), meningioma (17.3%), tumors of cranial and spinal nerves (12%), and metastatic lesions (12%). Diagnostic accuracy of squash cytology was higher in glial tumors (93.8%), meningioma (96.7%), and metastatic lesions (95.45%). Diagnostic accuracy of radiological modalities was 85.78%. Conclusions: A good familiarity with cytomorphological features of CNS lesions, clinical details, radiological findings, and intraoperative impression of neurosurgeon enables the pathologist to improve diagnostic accuracy and reduce errors.
Keywords: CNS, diagnostic accuracy, intraoperative, squash
How to cite this article:The incidence of CNS tumors has increased in recent years, and it is higher in developed countries (5.1 per 100,000) than in developing countries (3 per 100,000).[1] In India, incidence of CNS tumors ranges from 5 to 10 per 100,000 population.[2]
In 1920s, Eisenhardt and Cushing introduced intraoperative cytologic investigations for quick examination of neurosurgical specimens.[3]
The study aims to analyze the cytomorphological features of CNS lesions in intraoperative squash smears and to compare it with histopathology, immunohistochemistry, and preoperative radiological diagnosis. The diagnostic accuracy of intraoperative squash cytology in various CNS lesions was also determined.
Subjects and MethodsEthics committee approval was obtained and the approval certificate is being attached in the same mail. Date of approval is 1st November 2018. This was a prospective study conducted over a period of two years in a tertiary care centre. Central nervous system lesions which had both squash cytology and histopathology were included. A total of 190 cases were included in the study. The biopsy samples obtained at the time of surgery were transported immediately to the pathology laboratory in isotonic saline for processing. Smears were prepared by placing 1–2 mm of biopsy material at one edge of clean, dry, and labeled slide and smeared with another slide with just enough pressure to spread the tissue into thin film. Half of the smears made were air-dried and stained with Toluidene blue, and remaining smears were fixed in 100% isopropyl alcohol for quick staining with hematoxylin and eosin (H and E). After preparation of squash smears, the remaining tissue (if present) was fixed in 10% buffered formalin and processed for histopathologic evaluation.
The squash cytosmear and histopathological slides were reviewed by two different senior pathologists. Discordant cases were evaluated, and possible reason for same was recorded. Diagnostic difficulty encountered was overcome with aid of immunohistochemistry. The age, sex ratio, concordances, and discordances in diagnosis were analyzed. The cases were categorized into true positives, false positives, true negatives, and false negatives, and 2*2 table was made for each main category. Diagnostic accuracy was calculated in all CNS lesions. Sensitivity and specificity were calculated in lesions where cases could be categorized and arranged in 2*2 table.
ResultsA total of 190 cases were included in the present study. Thirty-eight cases (20%) belonged to pediatric age-group (0-12 years). A total of 152 cases (80%) of cases were adults. Maximum incidence was in age-group of 41–50 years or 5th decade of life.
One hundred and two cases (53.68%) were males, and 88 cases (46.32%) were females.
Among 190 cases, eight cases were non-neoplastic and 182 were neoplastic.
Diagnostic accuracy in non-neoplastic lesions was 88.8%. The sensitivity and specificity were 88% and 100%, respectively.
The percentage of various primary CNS neoplastic lesions is shown in [Figure 1].
The diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of various CNS lesions are given in [Table 1].
Table 1: Diagnostic accuracy, sensitivity, specificity, *positive predictive value, **negative predictive value of various primary CNS neoplastic lesionsA total of 184 cases had a specific preoperative diagnosis, and six cases were diagnosed as space occupying lesion (SOL). Discordance in diagnosis was noted in 27 cases, and diagnostic accuracy of preoperative radiological diagnosis was 85.78%.
DiscussionIntraoperative squash or crush technique is a simple, rapid, reproducible, and reliable technique giving good cytological details for arriving at accurate diagnosis.[4]
Touch preparations and smear preparations which are most commonly used are ideal for rapid intraoperative diagnosis of stereotactic brain biopsy specimens because the tissue is minimally distorted and none is wasted.[5]
Gliomas were the commonest CNS neoplasm in the study of 65 cases (35.7%). These cases were classified as low-grade gliomas (Grade I and Grade II lesions) and high-grade gliomas (Grade III and Grade IV).
Thirty-seven cases were histologically diagnosed as low-grade gliomas. Discordance was noted in five cases pertaining to typing of tumor and in four cases pertaining to grading of tumor.
Oligodendrogliomas were most often misdiagnosed as astrocytomas or mixed gliomas with probability of 69.2%, due to the lack of uniform-appearing nuclei and simulation of cytoplasmic processes.[6] Perinuclear halo characteristic of oligodendroglioma on histological preparation is a fixation artifact and therefore is not usually seen in cytological preparation.[7],[8] Presence of calcification in smear, mini-gemistocytes, sparse glial fibers, and perineuronal satellitosis favor diagnosis of oligodendroglioma.[9]
Ependymomas were misdiagnosed as choroid plexus papilloma on squash cytology.
Choroid plexus tumor can be confused with papillary ependymomas in squash cytology due to absence of gliofibrillary background, epithelial organization, and well-organized papillary architecture.[4]
Arrangement of tumor cells encasing blood vessels separated by perivascular cell-free zones of fibrillary processes (perivascular pseudorosette) is essential for the diagnosis of ependymomas on smears and helps in distinguishing it from ventricular or periventricular tumors.[10] Biopsy from an infiltrating margin and inadequate smearing were few of the limitations for accurate grading, and in such cases, only low-grade or high-grade astrocytoma could be suggested on smears.[4] Another reason for overdiagnosis or misdiagnosis in smears was intratumoral variability in the grading of the same tumor.[10],[11] Therefore, it is not advisable to grade gliomas based on small biopsies.[12]
Twenty-eight high-grade gliomas were present in study. Discordance pertaining to typing of tumor was seen in three cases [Figure 2]. Astrocytomas were misdiagnosed as ependymomas with a probability of 45.6% due to similarity between perivascular astrocytic processes and ependymal rosettes. The probability of misdiagnosing ependymomas as astrocytomas, mixed gliomas, or glioblastomas is 70.6% because of absence of ependymal rosettes.[6] Similar difficulties were encountered in our study and has led to discordances in same. Overall diagnostic accuracy of squash cytology in glial tumors is 93.8%.
Figure 2: High Grade Glioma with Giant cells, Correlation with histology and Ki67Among 31 meningiomas included in the study, difficulty was observed while typing of meningiomas into its variants by intraoperative cytology.
Fibrous meningioma was misdiagnosed as vascular or angiomatous lesion due to the lack of characteristic cytological features and due to large areas of hemorrhage.
Among the embryonal tumors, one case of anaplastic ependymoma was misdiagnosed as embryonal tumor in squash cytology. There is 11.7% probability of misdiagnosing ependymoma as a medulloblastoma.[6] Ependymomas which lacked true ependymal rosettes can lead to misdiagnosis as primitive neuroectodermal tumor. Ependymomas arising from posterior fossa location with presence of small cells and elongated nuclei can cause confusion with medulloblastoma. Another case of small cell glioblastoma multiforme with absence of fibrillary background was also misdiagnosed as medulloblastoma.[4]
Only two histopathologically confirmed cases of ganglioglioma were included in the study. One case of recurrent oligodendroglioma was misdiagnosed as ganglioglioma in squash cytology due to lack of perinuclear halo. Mini-gemistocytes in oligodendroglioma were mistaken for ganglion cells along with fibrillary matrix. Difficulties in distinguishing neuronal and mixed neuronal glial tumors from glial tumors were also recorded by Goel et al. and Perry et al.[4],[13]
Twenty-two metastatic lesions of CNS were included in the study, and one case showed discordance. Diagnostic difficulty was encountered in differentiating it from high-grade gliomas because of high cellularity, pleomorphism, and the round-to-polygonal appearance of the cells. There is a 42.9% probability of misdiagnosing glioblastoma with metastatic carcinoma and vice versa.[6] It was difficult to determine tumor differentiation of metastatic carcinomas on smear cytology.[9] Therefore, the diagnostic accuracy of squash cytology in diagnosis of metastatic tumors was calculated as 95.45%. Among germ cell tumors, histiocytic tumors and lymphoproliferative lesions' complete concordance was noted.
Diagnostic accuracy of preoperative radiological diagnosis based on CT and conventional MRI was calculated as 85.78% in the present study.
ConclusionSquash smear cytology is an accurate, reliable, rapid, safe, cost-effective, and intraoperative diagnostic method giving very good nuclear and cytoplasmic details of intracranial lesions in resource-limited settings. A good familiarity with the cytomorphological features of CNS lesions along with the clinical and radiological correlation can improve diagnostic accuracy of intraoperative squash cytology.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
Correspondence Address:
Dr. Shilpa Accamma Philip
Madappattil House, Kuttapuzha P.O., Tiruvalla - 689103, Kerala
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/joc.joc_70_22
留言 (0)