Graphical abstract

INTRODUCTION

Artificial intelligence (AI), defined as the simulation of human intelligence by machines, has become a transformative force in medicine, particularly in diagnostic disciplines such as pathology [1]. Historically, pathology has relied on light microscopy and manual interpretation of glass slides, making diagnostic processes labor-intensive and subjective. However, the advent of whole slide imaging (WSI) scanners has revolutionized the field by enabling digital pathology. WSI technology allows the digitization of glass slides into high-resolution images, providing enhanced opportunities for data storage, retrieval, and computational analysis that were not possible with traditional microscopy [2,3].

The transition from analog to digital workflows has been further supported by the widespread availability of high-speed internet, cloud storage, and a growing interest in telepathology, enabling remote diagnostic practices and collaboration. These advancements have laid the foundation for the integration of AI into pathology workflows, improving the precision and efficiency of diagnosis [4-6].

Among AI technologies, automated analysis of histopathological images, applications such as tumor detection, immunostaining evaluation, pattern recognition, and diagnostic classification have shown promising results. Additionally, AI tools have been employed for tasks requiring precision, such as tumor margin assessment and recommending diagnostic or genetic panels [7-11]. For example, based on morphological assessment, AI systems may recommend appropriate immunohistochemical stains or molecular/genetic panels to support accurate diagnosis and guide personalized treatment planning.

Despite its potential, AI adoption in pathology remains in its infancy, especially in resource-constrained settings like India, where challenges such as limited infrastructure, insufficient training, and lack of awareness persist [7,12]. Recognizing the importance of understanding AI’s role and adoption readiness, this study aims to assess pathologists’ knowledge, perceptions, and interest in AI applications. By analyzing responses from professionals with diverse backgrounds, the study identifies gaps and barriers to AI adoption while offering actionable insights for effective integration into pathology practices.

MATERIALS AND METHODS

Study design, study area, study duration and sampling technique

The cross-sectional study was conducted among pathologists from April 1, 2024 to June 30, 2024 who were working in northern India after obtaining approval from the Institutional Ethics Committee of Dr. Radhakrishnan Government Medical College, Hamirpur, Himachal Pradesh (IEC No. HFW-H-Dr. RKGMC/Ethics/2023/23 dated 10/20/23). Convenience sampling was used to collect the data from participants per the inclusion and exclusion criteria.

Inclusion criteria included (1) practicing pathologists (senior residents, faculty, private practitioners); (2) affiliated with academic institutions, tertiary care centers, or private labs; (3) willing to participate and located in Northern India.

Exclusion criteria were (1) non-pathologists or non-practicing pathologists and (2) incomplete or duplicate survey responses.

Data collection tool

Data were collected from the participants with the help of a structured and anonymous questionnaire to explore the attitudes of pathologists toward AI. The questionnaire was meticulously designed to capture a broad spectrum of information, including demographics, professional background, AI awareness, technological proficiency, and practical experiences with AI in pathology.

The questionnaire was divided into three key thematic sections and comprised 15 questions as shown in Fig. 1. The first section focused on demographics and professional background (Q1–Q5). This section gathered details such as age, sex, professional designation, years of experience, and practice setting (academic, private, or institutional). The second section addressed awareness and knowledge of AI (Q6–Q12), evaluating participants’ familiarity with AI concepts, sources of AI knowledge, self-assessed expertise, prior exposure to AI in pathology, and formal training. Participants rated their AI knowledge using four categories: ‘no knowledge’ (never heard of AI in pathology), ‘basic knowledge’ (general awareness without practical experience), ‘good knowledge’ (familiarity with concepts and some tool exposure), and ‘excellent knowledge’ (confidence in understanding and practical application). These were self-perceived levels, and no external validation was applied, which is an acknowledged limitation.

In the third section, technological proficiency and current practices (Q13–Q15) were assessed. This section assessed respondents’ comfort with digital tools, access to WSI, and their reliance on diagnostic modalities like traditional microscopy.

Data collection method

Data were collected by administering a questionnaire online using Google Forms. The link was disseminated electronically through email and professional WhatsApp groups targeting pathologists across various institutions in Northern India. Participation was entirely voluntary, and anonymity was ensured to encourage unbiased responses. Consent was obtained from the participants through a Google Form, and confidentiality of data was ensured.

Statistical analysis

Data were entered in Microsoft Excel 2010 (Microsoft, Richmond, WA, USA) and analyzed. The categorical variables were expressed as frequency and percentage. The continuous variables were expressed as mean and standard deviation. Comparison of categorical variables was carried out using chi-square test and Fisher’s exact test. All statistical analysis was done at a 5% level of significance, and a p-value < .05 was considered as significant.

RESULTS

Demographics and professional background

A total of 138 pathologists completed the survey. An overview of the age and sex distribution of the respondents as well as other demographic data are presented in Fig. 2. Detailed individual-level survey responses are provided in Supplementary Table S1.

The majority of respondents were female (66.7%) and aged 30–40 years (54.3%). Most participants were affiliated with academic institutions (78.3%), and faculty members with more than five years of experience constituted the largest subgroup (36.2%). The majority of participants were working in an academic environment, namely medical colleges (37.7%) and tertiary care centers (34.8%), with 21.7% of respondents working in private practice.

Awareness and knowledge of AI

The survey revealed high levels of awareness regarding AI’s role among pathologists, with 88.4% of respondents acknowledging its growing importance. Sex-based differences were found to be statistically significant, with females exhibiting greater awareness than males (93.5% vs. 78.3%, p = .008) as shown in Table 1.

Age-based differences were found to be statistically significant, with older respondents exhibiting greater awareness than younger respondents (95.8% vs. 84.4%, p = .047) as shown in Table 2.

Social media (44.2%) and conferences (41.3%) were the most frequently cited sources of AI knowledge. Formal training in AI was reported by only 6.5% of participants, underscoring a significant gap in education. When asked to self-assess their knowledge of AI in pathology, the majority reported basic knowledge (41.3%) or limited knowledge (39.1%), while only 5.8% rated their knowledge as excellent.

Academic pathologists were significantly more likely to have read medical publications about AI than their non-academic counterparts (57.4% vs. 26.7%, p = .003) as shown in Table 3. A similar trend was observed among females, with a near-significant difference in the number of publications read compared to males (p = .054), as shown in Table 1. Encouragingly, 92.8% of respondents expressed interest in attending AI-related workshops or courses, reflecting a readiness to bridge this knowledge gap.

Technological proficiency and AI application

Only 16.7% of respondents had used AI as a diagnostic tool in real-world pathology practice. Respondents with good or excellent knowledge were found to be significantly more likely to identify themselves as tech-savvy compared to those with basic or no knowledge (80.0% vs. 43.4%, p = .001), as shown in Table 4.

The respondents with good or excellent AI knowledge also reported significantly better access to WSI compared to those with limited knowledge (48.0% vs. 22.1%, p = .008).

DISCUSSION

This study provides valuable insight into the current state of AI awareness, knowledge, and adoption among pathologists in India. The high levels of awareness observed in this study align with global trends indicating growing interest in AI’s transformative potential for pathology [13,14]. However, significant gaps in formal training and practical application highlight barriers that must be addressed to fully realize AI’s benefits [15].

The influence of sex and professional setting on AI awareness is noteworthy. While female respondents demonstrated significantly higher awareness levels, this difference may be partially explained by professional affiliation. Notably, 82.6% (76 of 92) of female respondents worked at academic institutions, compared to 58.7% (27 of 46) of males. Given that academic environments offer greater access to AI-related resources and training, this distribution suggests a potential confounding effect, and the observed sex-based differences should be interpreted with caution. Additionally, respondents working in academic institutions exhibited greater familiarity with AI literature, likely due to better access to academic resources and exposure to research-oriented environments.

Although awareness of AI among pathologists was high, most respondents rated their knowledge as either basic or limited, highlighting a general lack of deeper understanding. The absence of formal training in AI further reinforces the need for well-structured educational programs to equip pathologists with necessary skills. On a positive note, the majority of participants expressed a strong interest in attending AI-related workshops or courses, showing a clear willingness to close this knowledge gap. These findings emphasize the need to incorporate AI training into pathology education, especially in countries like India, where the adoption of AI in practice is still at an early stage.

Additionally, a pattern emerged suggesting that access to advanced technologies like WSI and comfort with digital tools were key contributors to higher levels of AI knowledge. The limited use of AI tools in practice underscores challenges such as inadequate infrastructure, insufficient training, and limited access to enabling technologies like WSI. These findings suggest that digital pathology infrastructure is crucial for fostering familiarity and expertise with AI-driven workflow.

Interestingly, self-identified tech-savvy respondents demonstrated a stronger association with higher levels of AI knowledge, highlighting the importance of digital literacy in adapting to emerging technologies. These findings underscore the need to prioritize technological training and infrastructure development to promote equitable AI adoption across diverse practice settings.

This study highlighted notable differences in AI exposure between academic and private practice settings, with respondents in academic environments demonstrating greater familiarity with AI literature and concepts. This disparity likely stems from variation in resource availability, institutional support, and exposure to research-oriented activities.

The challenges identified in this Indian cohort mirror those observed globally. Worldwide, the integration of AI into pathology has been met with enthusiasm but also with considerable barriers. A bibliometric analysis revealed that the United States leads the field of AI-based tumor pathology research, contributing 41.3% of publications, followed by China and the United Kingdom [16,17]. This underscores the significant investment and focus on AI in pathology within these countries. In Europe, initiatives such as the Ecosystem for Pathology Diagnostics with AI Assistance (EMPAIA) project have been instrumental in accelerating AI adoption. EMPAIA has facilitated open-source platforms and interoperability standards, facilitating the integration of AI into clinical workflows across multiple laboratories [18]. Despite such efforts, Chua et al. [19] described system-wide challenges to AI implementation in oncology, including data bias, dynamic knowledge, and user interface design, and proposed actionable steps for overcoming these barriers through multidisciplinary collaboration. These insights affirm that the limitations seen in our study are not isolated but part of a broader global picture.

Infrastructural and technological barriers remain central to the slow uptake of AI in pathology. Specific challenges include the high cost of digital pathology systems, variable internet connectivity, limited access to computational resources, and a lack of trained personnel to manage and interpret AI outputs. Additionally, digital literacy levels vary significantly, particularly in private practice or rural settings. Data security concerns, particularly in handling sensitive patient images and records, further complicate the deployment of AI systems. Addressing these multi-layered issues will require a coordinated response involving institutional investment, government support, and public–private partnerships to create a conducive ecosystem for AI adoption.

One observation from our study is the overrepresentation of younger pathologists, with 65.2% under the age of 40. Notably, this age distribution is consistent with trends reported in both Indian and global pathology surveys [20,21]. This likely reflects greater digital engagement among younger professionals, which may introduce bias and limit the generalizability of the findings to the broader population of practicing pathologists. Future studies involving a more geographically diverse cohort of pathologists would enhance the statistical robustness of the findings and offer a more comprehensive understanding of AI awareness and adoption across different regions.

In conclusion, while awareness and interest in AI are growing among Indian pathologists, there remains a clear gap between recognition and implementation. To bridge this divide, actionable steps must include the development of standardized AI curricula for pathologists; organizing national and regional workshops; fostering collaborations between academic institutions, private sector developers, and healthcare policymakers; and expanding access to digital infrastructure and secure data-sharing platforms. By investing in education, policy, and technology together, the pathology community in India—and globally—can accelerate meaningful and equitable integration of AI into clinical practice.

Supplementary Information

Fig. 1.

Survey on artificial intelligence (AI) in pathology.

Fig. 2.

Demographic details of respondents. (A) Age distribution of respondents. (B) Sex distribution of respondents. (C) Professional background of respondents.

REFERENCES

• 1. Tizhoosh HR, Pantanowitz L. Artificial intelligence and digital pathology: challenges and opportunities. J Pathol Inform 2018; 9: 38.ArticlePubMedPMC
• 2. Bauer TW, Schoenfield L, Slaw RJ, Yerian L, Sun Z, Henricks WH. Validation of whole slide imaging for primary diagnosis in surgical pathology. Arch Pathol Lab Med 2013; 137: 518-24. ArticlePubMedPDF
• 3. Fertig RM, Sangueza O, Gaudi S, Gamret AC, Cervantes J, Jukic DM. Whole slide imaging. Am J Dermatopathol 2018; 40: 938-9. ArticlePubMed
• 4. Evans AJ, Salama ME, Henricks WH, Pantanowitz L. Implementation of whole slide imaging for clinical purposes: issues to consider from the perspective of early adopters. Arch Pathol Lab Med 2017; 141: 944-59. ArticlePubMedPDF
• 5. Hartman DJ, Pantanowitz L, McHugh JS, Piccoli AL, MJ OL, Lauro GR. Enterprise implementation of digital pathology: feasibility, challenges, and opportunities. J Digit Imaging 2017; 30: 555-60. ArticlePubMedPMCPDF
• 6. Thrall M, Pantanowitz L, Khalbuss W. Telecytology: clinical applications, current challenges, and future benefits. J Pathol Inform 2011; 2: 51.ArticlePubMedPMC
• 7. Bera K, Schalper KA, Rimm DL, Velcheti V, Madabhushi A. Artificial intelligence in digital pathology: new tools for diagnosis and precision oncology. Nat Rev Clin Oncol 2019; 16: 703-15. ArticlePubMedPMCPDF
• 8. Echle A, Rindtorff NT, Brinker TJ, Luedde T, Pearson AT, Kather JN. Deep learning in cancer pathology: a new generation of clinical biomarkers. Br J Cancer 2021; 124: 686-96. ArticlePubMedPDF
• 9. Couture HD, Williams LA, Geradts J, et al. Image analysis with deep learning to predict breast cancer grade, ER status, histologic subtype, and intrinsic subtype. NPJ Breast Cancer 2018; 4: 30.ArticlePubMedPMCPDF
• 10. Lu MY, Chen TY, Williamson DF, et al. AI-based pathology predicts origins for cancers of unknown primary. Nature 2021; 594: 106-10. ArticlePubMedPDF
• 11. Strom P, Kartasalo K, Olsson H, et al. Artificial intelligence for diagnosis and grading of prostate cancer in biopsies: a population-based, diagnostic study. Lancet Oncol 2020; 21: 222-32. ArticlePubMed
• 12. Madabhushi A, Lee G. Image analysis and machine learning in digital pathology: challenges and opportunities. Med Image Anal 2016; 33: 170-5. ArticlePubMedPMC
• 13. Baxi V, Edwards R, Montalto M, Saha S. Digital pathology and artificial intelligence in translational medicine and clinical practice. Mod Pathol 2022; 35: 23-32. ArticlePubMedPDF
• 14. Drogt J, Milota M, Vos S, Bredenoord A, Jongsma K. Integrating artificial intelligence in pathology: a qualitative interview study of users' experiences and expectations. Mod Pathol 2022; 35: 1540-50. ArticlePubMedPMCPDF
• 15. Sajithkumar A, Thomas J, Saji AM, et al. Artificial intelligence in pathology: current applications, limitations, and future directions. Ir J Med Sci 2024; 193: 1117-21. ArticlePubMedPDF
• 16. Shen Z, Hu J, Wu H, et al. Global research trends and foci of artificial intelligence-based tumor pathology: a scientometric study. J Transl Med 2022; 20: 409.ArticlePubMedPMCPDF
• 17. Hanna MG, Pantanowitz L, Dash R, et al. Future of artificial intelligence-machine learning trends in pathology and medicine. Mod Pathol 2025; 38: 100705.ArticlePubMed
• 18. Zerbe N, Schwen LO, Geissler C, et al. Joining forces for pathology diagnostics with AI assistance: the EMPAIA initiative. J Pathol Inform 2024; 15: 100387.ArticlePubMedPMC
• 19. Chua IS, Gaziel-Yablowitz M, Korach ZT, et al. Artificial intelligence in oncology: path to implementation. Cancer Med 2021; 10: 4138-49. ArticlePubMedPMCPDF
• 20. Kanungo R, Sharma A, Pagi S, Sachar K. Exploring occupational health challenges in pathology: a cross-sectional survey of Indian pathologists. Cureus 2024; 16: e75264. ArticlePubMedPMC
• 21. Sarwar S, Dent A, Faust K, et al. Physician perspectives on integration of artificial intelligence into diagnostic pathology. NPJ Digit Med 2019; 2: 28.ArticlePubMedPMCPDF

Figure & Data

References

Citations

Citations to this article as recorded by