|Year : 2012 | Volume
| Issue : 3 | Page : 183-189
|Breast fine needle aspiration cytology practices and commonly perceived diagnostic significance of cytological features: A pan- India survey
Hrushikesh Tukaram Garud1, Debdoot Sheet2, Manjunatha Mahadevappa2, Jyotirmoy Chatterjee2, Ajoy Kumar Ray3, Arindam Ghosh4
1 School of Medical Science and Technology, Indian Institute of Technology Kharagpur, Kharagpur, Video Signal Processing, Texas Instruments (India) Pvt. Ltd., Bangalore, India
2 School of Medical Science and Technology, Indian Institute of Technology Kharagpur, Kharagpur, India
3 Department of Electronics and Electrical Communication Engineering, Indian Institute of Technology, Kharagpur; Bengal Engineering and Science University, Shibpur, Howrah, India
4 Department of Pathology Sub-Divisional Hospital, Kharagpur, India
Click here for correspondence address and email
|Date of Web Publication||21-Sep-2012|
| Abstract|| |
Background: Methodical and meticulous understanding of clinico-pathological procedures and decision making process of cancer diagnosis and identification of aspects that are well-suited for computer-aided analysis are first steps toward development of assistive computational tool for analysis of breast fine needle aspiration cytology (FNAC) slides.
Aims: To identify variables in practice of FNAC as used for diagnosis of breast lesions and commonly perceived diagnostic significance of cytological features for diagnosis of benign or malignant condition of breast lesions.
Materials and Methods: An India-wide questionnaire-based survey of cytopathologists/pathologists' breast FNAC reporting practices and their opinion on diagnostic significance of cytological features in diagnosis of benign or malignant nature of breast lesion were conducted.
Results: Fifty-one experts working with various medical education institutes (~52% of participants), oncological tertiary care centers (~28%) and primary care centers/private diagnostic pathology laboratories (~20%) spread over 13 states of India have participated in the survey. Constants and variables observed in clinico-cytopathological practices and combined opinion of the participants on diagnostic significance of cytological features are presented here.
Conclusions: There exist analogous as well as varied components in clinico-pathological procedures and diagnostic interpretation by individuals. These constants and variables in the practice of breast FNAC should be considered, when drawing up specifications for an assistive computational tool for analysis of breast FNAC slides. The estimate for commonly perceived significance of cytological features obtained through this study will help in their selection for computer-aided analysis of breast FNAC slides and further in selection of corresponding feature quantification techniques.
Keywords: Breast; clinical practices; FNAC; survey
|How to cite this article:|
Garud HT, Sheet D, Mahadevappa M, Chatterjee J, Ray AK, Ghosh A. Breast fine needle aspiration cytology practices and commonly perceived diagnostic significance of cytological features: A pan- India survey. J Cytol 2012;29:183-9
|How to cite this URL:|
Garud HT, Sheet D, Mahadevappa M, Chatterjee J, Ray AK, Ghosh A. Breast fine needle aspiration cytology practices and commonly perceived diagnostic significance of cytological features: A pan- India survey. J Cytol [serial online] 2012 [cited 2020 Jun 3];29:183-9. Available from: http://www.jcytol.org/text.asp?2012/29/3/183/101168
| Introduction|| |
Fine needle aspiration cytology (FNAC) is a part of triple assessment of breast lesions and has been recognized as the most cost-effective procedure with shortest turnaround time.  In current practice of breast FNAC, diagnosis is based on subjective assessment of microscopic appearance of the aspirates, therefore, difficulties in maintaining consistency and reproducibility in findings are inevitable. Moreover, there exists an overlap in cytological features of benign and malignant lesions which may lead to equivocal diagnosis.  When using five reporting categories for breast cytology (i.e., C1-inadequate; C2-benign; C3- atypical; C4-suspicious; and C5-malignant),  atypical and suspicious findings indicate equivocal diagnosis wherein tissue biopsy is necessary for definite diagnosis. These findings combined together have reported rates of 6.9-20%,  whereas, Al-Kaisai  found that the true gray zone in breast cytology is 2% of all cases. Thus there is a scope to reduce the rate of equivocal diagnosis.
Also there exists differential preference for clinico-cytopathological procedures and discordance among expert cytopathologists/pathologists in morphometric diagnosis. Pathologists either do not use the same criteria for diagnosis or apply the same criteria differently from one another.  Alternatively it has been observed that globally cytopathologists/pathologists see the same but locally they see different things. 
To overcome the inherent limitations presented by subjective evaluation of cytological smears and improve the reliability of diagnosis, it is important to develop and use assistive computational tools for breast cancer diagnosis that operate on objective features and quantitative measures. In literature, many such computational tools for breast cancer diagnosis can be found with most of them concentrating on classification of FNAC slides as benign or malignant and some focusing on staging of malignancy.  When designing and developing such computational aid it is necessary to understand practices and perceptions of the experts to identify aspects suitable for computer-aided analysis. This will also help define user centric functionality aspects of the system. In this regard a nationwide questionnaire-based survey titled 'Practice of FNAC as a screening tool in clinical management of symptomatic patients of breast cancer' was conducted. The findings of the survey are reported here.
| Materials and Methods|| |
The survey was designed with primary goals to
- Identify common practices and variables in clinico-cyopathotological procedures of aspiration, sample processing and visual assessment using microscope in reporting of breast FNAC. This will help define user centric functionality aspects of the system
- Identify commonly perceived diagnostic significance of cytological features for diagnosis of benign or malignant condition of breast lesions. This will help develop better pattern recognition algorithms.
To complement these goals, the questionnaire for pathologist comprised of two parts. First part of the questionnaire included questions on clinico-cytopathological procedures such as sample collection and processing techniques used, type of stains used and additional investigations performed routinely in breast FNAC.
In the second part of the questionnaire, participants were asked to specify the importance level they attach to each cytological feature during diagnosis of benign or malignant conditions. For each cytological feature, the participants had to select one of the five predefined linguistic significance levels - i) Not significant (0), ii) Slightly significant (1), iii) Moderately significant (2), iv) Significant (3) and v) Most significant (4). Numerical values in the brackets represent corresponding numerical values used to combine individual opinions and find common perception. Simultaneously, participants also had to specify the microscope objective magnification they preferably use while assessing condition of cytological features. Participants were also requested to provide details about their reporting experience and personal confidence level on FNAC as a screening tool in management of symptomatic patients of breast cancer. Help of scholarly texts in clinical pathology such as 'Manual and Atlas More Details of Fine Needle Aspiration Cytology',  'Fine Needle Aspiration Cytology: Diagnostic Principles and Dilemmas',  Cytology: Diagnostic Principles and Clinical Correlates  and 'Textbook of Pathology'  was taken while preparing the questionnaire.
Participation in the survey was voluntary and questionnaire booklet carried 'Statement of Purpose' which apprised participants about purpose of the survey. Hundred questionnaires were sent to the heads of the pathology/cytology departments at renowned medical education and research institutes, oncological tertiary care centres and primary care centres/private diagnostic pathology laboratories (total 50 institutes). The head of the departments (institutes) were requested to distribute the questionnaires within their institutes to the concerned cytology/pathology experts they felt could best provide the representative information.
| Results of the Survey|| |
Currently we have received responses from 51 cytopathologists/pathologists from renowned medical education institutes (52% of the responses received), oncological tertiary care centres (28%) and primary care centres/private diagnostic pathology laboratories (20%) spread over 13 states of India (Assam, Andhra Pradesh, Bihar, Chandigarh, Karnataka, Madhya Pradesh, Maharashtra, Punjab, New Delhi, Orissa, Rajasthan, Tamil Nadu and West Bengal). Individual reporting experiences of the participants are in the range of 2-41 years with average reporting experience of 17 years.
Here findings of the survey are reported in two parts. Part-I presents findings of the survey related to breast FNAC procedures, which include tissue sample collection, processing, staining techniques and immunocytochemical investigations used by cytopathologists/pathologists. Part-II presents findings on commonly perceived significance level of the cytological criteria in diagnosis of benign or malignant condition of breast lesions.
Part I: The constants and variables in breast FNAC practices
Most of the participants of the survey use different sizes of the needles for aspiration varying from case to case, with most of them using finer needles (>22 G). Approximately 62% of the participants of this survey use 23G needle, while 39% of the participants prefer 22-G needles [Table 1].
|Table 1: Usage data for different needle sizes in breast FNAC aspiration|
Click here to view
Image guidance for aspiration
Image guidance during aspiration of breast lesions is helpful for exact localization of the lump. Overall 74% percent of the participants use image guidance for aspiration. However, image guidance might not be necessary for aspiration of obvious and palpable or superficial and definite lesions, which is evident, as 62% of the participants have reported using it only sometimes (not for every aspiration performed [Table 2]).
Seventy-nine percent of the participants using image guidance prefer ultrasound. Availability, ease of use, cost of ownership and operation, and real-time intervention capabilities seem to influence choice of the imaging modality for guided aspiration. The usage data for different imaging modalities is reported in [Table 3].
|Table 3: Usage data of imaging modalities for guidance during breast lump aspiration|
Click here to view
Fixation and staining methods used for primary diagnosis from FNAC
Most number of participants of the survey report using both wet-fixation and air-drying fixation techniques in slide preparation [Table 4], whereas three staining methods are routinely used in breast primary cyto-diagnosis, namely Romanowsky, hematoxylin and eosin (H and E) and Papanicolaou (Pap). However, every cytopathologist/ pathologist seems to favor the method that was used in his/her training institute. Most of the participant cytopathogists/pathologists, in this survey, use more than one slide prepared from the aspirate of a single lump. Use of both wet-fixed and air-dried slides for each aspirate is done by 66% of the participants [Table 5]. It has also been observed that the use of only air-drying fixation during slide preparation is more common amongst the participants from primary care centers/ private pathology laboratories as compared to participants from medical education institutes or tertiary care centers.
|Table 4: Usage data for tissue fixation techniques in breast FNAC slide preparation for primary diagnosis|
Click here to view
|Table 5: Staining methods used in breast FNAC slide preparation for primary diagnosis|
Click here to view
Additional staining methods
In addition to routine staining combinations, more types of staining techniques [Table 6] are used by 20% of the participants. Sixty percent of the participants using additional stains are from medical education institutes and remaining from primary or tertiary care centers. Stains used for confirmation of mucinous carcinoma, are the most commonly used additional stains.
Immunocytochemical investigations are performed by 27% of the participants of this survey; 71% of the participants performing immunocytochemical investigations are from tertiary care centers and rest 29% from educational institutions. None of the participants from primary care centers/private pathology laboratories perform immunocytochemical investigations. Almost 87% of the participants from tertiary care centers practice immunocytochemistry, as compared to 15% from educational institutes. Prognostic investigations such as check for estrogen progesterone (ER/PR) receptor status and human epidermal growth factor receptor-2 (HER-2/neu or c-erb-B2) status are the most commonly used immunocytochemical investigations practiced by 24% and 14% of the overall participants, respectively. Use of tumor markers such as cytokeratin (CK), vimentin (VIM), neuron-specific enolase (NSE), prostate-specific antigen (PSA), cathepsin-D (CATH-D), leukocyte common antigen (LCA) and p53 has also been reported by the participants. The usage data for different immunocytochemical investigations in breast FNAC is given in [Table 7].
Part II: Commonly perceived diagnostic significance of the cytological criteria
This work has been initiated with the belief that commonly perceived diagnostic significance of a cytological criterion represents its real ability in discriminating benign and malignant condition of a breast lump. Participants' percentage opinion is given in [Figure 1]. In order to combine individual opinions and find common perception of the participants mean shift technique  has been used. When combining available individual opinions three experiments were performed. Numerical findings of these experiments are presented [Table 8].
In first experiment, opinions of all the participants were combined to compute overall mean significance score (commonly perceived significance) for the features. Also, the coefficient of variation about the mean was computed to estimate the variability in opinion [Figure 2]. The features are ordered in nonincreasing fashion of overall mean significance scores.
|Figure 2: Overall opinion and the extent of variation among participants|
Click here to view
From overall mean significance scores it can be observed that 'Nuclear chromatin' is the only most-significant feature; 'nature of background', and 'volume and color of cytoplasm' are in moderately significant category, whereas, all other cytological features belong to the significant category. From [Figure 1] and [Figure 2] it can also be observed that there exists a considerable degree of concordance among the participants about diagnostic significance of most of the cytological criteria. Only in the case of four criteria, namely 'size and shape of aggregates', 'nature of background', 'number of nucleoli', and 'volume and color of cytoplasm' participants exhibit relatively higher degree of disagreement with coefficient of variation  of more than 0.3.
In second experiment, the responses were divided in three groups as from primary care centers/private pathology labs, medical education institutions and tertiary care centeres, and groupwise collective significance score were computed [Figure 3]. The overall mean significance scores have also been plotted for comparative evaluation. Perceived significance plots show high degree of disparity among three hierarchical levels of healthcare delivery, with collective opinion of the participants from medical education institutions closely following the trends of the overall mean opinion. The collective opinion of the participants from primary care centres/private pathology labs shows high dynamism and divergence from the overall mean opinion.
|Figure 3: Perceived significance of cytological features at different hierarchical levels of the healthcare delivery|
Click here to view
In experiment three, variations in perception of experts according to their reporting experience was studied. The responses were divided in three groups according to the reporting experience in years, as below 5 years, more than 5 years but less than average reporting experience (17 years) and more than average experience groupwise collective significance score were computed. Perceived significance scores have been plotted along with overall mean significance scores [Figure 4]. This figure shows converging trend in perceived significance levels with increasing experience, also difference of opinion with the overall opinion decreases with increasing reporting experience.
|Figure 4: Perceived significance of cytological features and effect of experience on it|
Click here to view
| Discussion|| |
Among the cytopathologist/pathologists concurrence is seen on the type of needles and image guidance to be used for aspiration of breast lumps. Interestingly more than half of the participants report of 'sometimes' using image guidance for aspiration; they otherwise perform freehand procedures. This might be attributed to a possible reflection of the scenario where late presentation of the disease with obvious and palpable lumps is common. Also, when choosing the imaging modality for guided aspiration, availability, ease of use, cost of ownership and operation, and real-time intervention capability of the modalities seem to influence the decision.
The findings also indicate highest degree of variation in practice of spreading and staining techniques used for primary diagnosis. Although most of the surveyed respondents agree on use more than one slides prepared from the aspirate of a single lump; a cytopathologist/pathologist mostly favor the staining method that was used at his/her training institute. It is also seen that no more than 25% participants use same stain combinations. Additional diagnostic/prognostic investigations like special staining and immunocytochemical investigations show high degree of divergence among different healthcare delivery levels. These investigations are not performed by any of the participants from the primary care centres/private pathology labs; only participants from medical education institutes and tertiary care centers report of performing them. Additional staining techniques are predominantly used by medical education institutes while immunocytochemical investigations are mostly performed at tertiary care centres.
Overall perception on the significance level of the cytological features in diagnostic interpretation of breast FNAC slides can be considered to represent the actual discrimination ability of the feature. The collective opinion shows disparity from one level of healthcare delivery model to the other. From the findings of the survey it can been observed that nuclear features like 'nuclear chromatin', 'prominence of nuclei', 'nuclear membrane', 'mitotic figure' are considered more significant by participants from medical education institutes and tertiary care centers; nuclear features and gross cytoarchitectural patterns such as 'size and shape of aggregates', 'cohesiveness of clusters' and 'presence of bare nuclei' are favored by participants from primary care centers; other cytological features viz., 'myoepthelial cell population', 'nuclear cytoplasmic ratio' and 'cellular pleomorphism' are preferred equally by all the participants. Also as observed earlier, the perception about the features shows converging trend with increasing experience. The perceived significance varies widely among the different experience groups for features such as 'cohesiveness of epithelial cells', 'number of nucleoli', 'nature of background', and 'volume and color of cytoplasm'; whereas the perceived significance of features 'nuclear cytoplasmic ratio', 'mitosis', 'bimodal pattern of aggregates', 'nuclear shape' and 'nuclear size' is almost similar for all experience groups.
| Conclusions|| |
From the nationwide survey on practice of FNAC as a screening/diagnostic tool in breast cancer diagnosis, it has been observed that there exist analogous as well as varied components in clinico-pathological procedures and diagnostic interpretation by individuals. These constants and variables in practice of FNAC for breast cancer screening/diagnosis should be considered, when drawing up specifications for an assistive computational tool for analysis of breast FNAC slides. Such a computational tool using objective features will ensure provision of relevant, appropriate and accurate assistance in diagnostic interpretation and its subsequent broad acceptability. The estimate for commonly perceived significance of cytological features obtained through this study will help in selection of cytological features and selection of corresponding objective feature quantification techniques to be used in analysis of breast FNAC slides.
| Acknowledgments|| |
We are grateful of all the institutes and individuals who kindly responded to our call and participated in this survey. We are also thankful to our sponsors Texas Instruments, Authorities at SMST, IIT Kharagpur and all those who directly or indirectly supported us in this activity.
| References|| |
|1.||Masood S, Vass L, Ibarra JA, Ljung BM, Stalsberg H, Eniu A, et al. Breast pathology guideline implementation in low-and middle-income countries. J Cancer 2008;113:2297-304. |
|2.||Kocjan G. Diagnostic dilemmas in FNAC: Difficult breast lesions. In: Fine needle aspiration cytology: Diagnostic principles and dilemmas. Berlin: Springer; 2006. p. 214-23. |
|3.||Wilson R, Asbury D, Cooke J, Michell M, Patnick J, editors. Clinical guidelines for breast cancer screening assessment. Sheffield (UK): NHS Cancer Screening Programmes; 2001. NHSBSP Pub. No 49. |
|4.||al-Kaisi N. The spectrum of the "gray zone" in breast cytology. A review of 186 cases of atypical and suspicious cytology. Acta Cytol 1994;38:898-908. |
|5.||Ackerman AB. Discordance among expert pathologists in diagnosis of melanocytic neoplasms. Hum Pathol 1996;27:1115-6. |
|6.||Crus-Ramírez N, Acosta-Mesa HG, Carrillo-Calvet H, Nava-Fernández LA, Barrientos-Martínez RE. Diagnosis of breast cancer using Bayesian networks: a case study. Comput Biol Med 2007;37:1553-64. |
|7.||Demir C, Yener B. Automated cancer diagnosis based on histopathological images: Asystematic survey. Technical Report TR-05-09. Troy (New York): Rensselaer Polytechnic Institute; 2005. |
|8.||Orell SR, Sterrett GF, Walters MN, Whitaker D, Lindholm K. Breast. In: Orell SR, Sterrett GF, Walters MN, Whitaker D, editors. Manual and atlas of fine needle aspiration cytology, 3rd ed. London: Churchill Livingstone; 1999. p. 145-200. |
|9.||Ducatman BS, Wang HH. Breast. In: Chibas ES, Ducatman BS, editors. Cytology: Diagnostic principles and clinical correlates, 3rd ed. Philadelphia: Saunders Elsevier; 2009. p. 221-54. |
|10.||Sharma R. Diagnostic cytopathology. In: Mohan H, editor. Textbook of pathology, 4th ed. New Delhi: Jaypee; 2002. p. 891-907. |
|11.||Cheng Y. Mean shift, mode seeking, and clustering. IEEE Trans. Pattern Anal. Mach Intell. 1995;17:790-9. |
|12.||McDonald JH, editor. Handbook of biological statistics, 2nd ed. Baltimore (Maryland): Sparky House Publishing; 2009. |
Hrushikesh Tukaram Garud
School of Medical Science and Technology, Indian Institute of Technology Kharagpur, Kharagpur-721302
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7], [Table 8]
|This article has been cited by|
||Comparative assessment of CNN architectures for classification of breast FNAC images
| ||Amartya Ranjan Saikia,Kangkana Bora,Lipi B. Mahanta,Anup Kumar Das |
| ||Tissue and Cell. 2019; 57: 8 |
|[Pubmed] | [DOI]|
| Article Access Statistics|
| Viewed||4659 |
| Printed||79 |
| Emailed||0 |
| PDF Downloaded||288 |
| Comments ||[Add] |
| Cited by others ||1 |