Application and exploration of digital pathology technology in standardized residency training

Abstract

Objective

To evaluate the effectiveness of digital pathology technology in the standardized residency training (SRT) in clinical pathology, compare the performance differences between digital slide-based and traditional microscope-based assessments, examine the association between assessment modality and residents’ slide-reading scores, explore residents’ perceptions and acceptance of digital pathology, and explore the potential presence of the “novice tool dependence” effect among junior residents.

Methods

A retrospective analysis was performed on assessment data from 64 resident physicians at the clinical pathology training base of Sir Run Run Shaw Hospital, Zhejiang University School of Medicine, between 2021 and 2024. Assessment scores from traditional glass slides (2021–2022) were compared with those from digital slides (2023–2024). Additionally, a questionnaire survey was conducted among 59 resident physicians from six SRT bases in Zhejiang Province to evaluate their usage, functional experience, and acceptance of digital slides. Open-ended responses were independently reviewed by two investigators. Similar responses were categorized into descriptive themes, and the frequency of each theme was documented. Since the open-ended responses were used to supplement and support the quantitative findings, no formal qualitative approach was applied.

Results

No significant difference was observed in assessment scores between the digital slide group (n = 33, 73.80 ± 11.34) and the traditional glass slide group (n = 31, 74.65 ± 13.45). Hierarchical multiple linear regression showed no statistically significant independent association between assessment modality and assessment scores after adjustment for residency year and educational background (β=-0.015, P = 0.878). In contrast, residency year (β = 0.607, P < 0.001) and educational background (β = 0.218, P = 0.030) were independently associated with assessment scores. Questionnaire responses indicated that the overall use rate of digital slides among resident physicians was 66.10%, with teaching and assessment as the most common applications (86.44%). Exploratory factor analysis showed that junior resident physicians scored significantly higher in the “diagnostic operation efficiency” dimension of digital slides (F = 4.614, P = 0.014). We tentatively describe this pattern as “novice tool dependence”. The Kruskal–Wallis H test showed relatively lower acceptance of digital slides for routine reading among senior residents compared to juniors (H = 6.636, P = 0.036). Open-ended responses suggested that residents generally perceived benefits of digital pathology in overcoming time and space limitations, enhancing learning motivation, and standardizing teaching resources across multiple hospital campuses. Nonetheless, participants reported concerns regarding system stability, image resolution, and network dependence.

Conclusion

Based on this single-center before-and-after comparison, digital slide-based assessment was not associated with a statistically significant difference in archived overall slide-reading scores. These findings provide preliminary support for the feasibility of digital slide-based assessment in SRT, but should be interpreted cautiously because of temporal confounding, limited sample size, and regional restrictions. The observed “novice tool dependence” among junior residents could serve as a preliminary reference for designing targeted SRT training curricula. Within the scope of the present sample, digital slides may serve as an auxiliary educational option for residency teaching and assessment; given limitations, including a limited sample size and regional restrictions, multi-center, cross-regional research is still needed to verify their applicability for wide-scale promotion.

Introduction

Progress in information technology and whole-slide imaging (WSI) has facilitated the shift from conventional pathology practice toward complete digital pathology (CDP) departments [1]. With the gradual integration of digital slides into routine microscopic diagnosis, relevant reform is underway within clinical pathology standardized residency training. In Zhejiang Province of China, digital slides have been officially applied for pathology resident graduation assessments starting in 2025. However, the suitability of this transition remains debated among experts, and most residents adapt to the updated assessment mode in a passive manner. Additionally, existing research has rarely systematically explored how digital slides influence examination outcomes, alongside residents’ acceptance and expectations toward this technique [2, 3]. Therefore, this study was designed to investigate the above concerns and generate preliminary reference data to inform the optimization of digital teaching and assessment strategies in pathology residency programs.

Methods

Study design

This study adopted a mixed-methods design, combining a retrospective cohort analysis with a prospective questionnaire survey. By analyzing assessment data and survey responses, we sought to preliminarily assess the application of digital slides in residency training from both objective performance and subjective perception perspectives.

Study participants

Retrospective analysis

Assessment data were collected from resident physicians at our clinical pathology training base, with the rollout of digital slide-based assessment in 2023 set as the cut-off time point. Data were included for two years before and after the reform. Inclusion criteria were residents who had completed at least one year of training and had fully recorded annual slide-reading assessment results. Exclusion criteria were absence from the assessment, make-up examinations, or incomplete records. A total of 64 annual assessment records met the inclusion criteria, comprising 31 cases in the traditional glass slide group (2021–2022) and 33 cases in the digital slide group (2023–2024). Since digital slide assessment was launched alongside routine institutional curriculum updates in 2023, this before-and-after comparison is non-randomized and may be affected by temporal confounding; accordingly, parallel random grouping was not feasible for the present analysis.

Prospective survey

Resident physicians from six SRT bases in Zhejiang Province were invited to complete an online questionnaire via the Wenjuanxing platform. Inclusion criteria included residents currently enrolled in training who voluntarily completed the questionnaire. Exclusion criteria covered graduates or trainees who had withdrawn, or responses with evident logical inconsistencies or invalid entries. A total of 59 valid questionnaire responses were obtained.

Study protocol

Slide reading assessment

The types of assessment specimens in the retrospective analysis complied with the requirements of Content and Standards for Standardized Residency Training (2022 Edition), and the scoring criteria followed Assessment Regulations for Clinical Practical Ability in the Qualification Assessment of Standardized Residency Training in Zhejiang Province. The traditional group adopted a microscope-based glass slide assessment, while the digital group used a whole-slide imaging (WSI) assessment. Examination item quantity, difficulty setting, and scoring rubrics were kept identical between groups. Hierarchical multiple linear regression was used to explore the potential association of assessment modality with scores. The archived assessment score corresponded to the overall composite score of the institutional slide-reading assessment within the SRT framework. The assessment primarily measured residents’ integrated performance in morphologic recognition, diagnostic interpretation, and case-based analytical reasoning according to institutional and provincial assessment standards. Detailed component-level subscores, such as separate scores for diagnostic accuracy, descriptive morphology, or differential diagnosis, could not be extracted from preserved historical archived dataset. Therefore, the present analysis was limited to focusing only on the overall composite assessment score.

Questionnaire survey

The questionnaire was developed specifically for this study based on three sources: published literature on digital pathology implementation and pathology education, practical experience from pathology SRT and digital slide assessment, and expert discussions among pathology educators and residency training supervisors. An initial item pool was generated to cover digital slide use frequency, application scenarios, perceived impact on assessment, functional and educational value, perceived advantages and limitations, and acceptance of different digital slide applications. The draft questionnaire was reviewed by three senior pathology faculty members with experience in residency education and digital pathology implementation. They assessed the relevance, clarity, and comprehensiveness of the items. Based on their feedback, wording was refined before formal distribution. Content validity was therefore partially supported through expert review, while construct validity was assessed using exploratory factor analysis. The survey was designed and distributed online via Wenjuanxing, covering aspects such as digital slide usage experience, perceived advantages and limitations, and acceptance (Table 1). The questionnaire showed good reliability and validity (Cronbach’s α = 0.813, KMO = 0.826, Bartlett’s test of sphericity P < 0.001), which allowed for exploratory factor analysis (EFA). Two factors were extracted and named “collaborative management efficacy” and “operational diagnostic efficacy” according to the conceptual meaning of the items with the highest factor loadings. Two pathology educators reviewed factor naming to ensure interpretability in the SRT context. Factor scores were tested for normality (Shapiro–Wilk), and homogeneity of variance (Levene’s test), followed by selection of appropriate parametric or non-parametric tests to compare group differences.

Table 1.

Key survey items used to evaluate digital slide use among resident physicians

Dimension	Item	Response Options
Frequency of Use	How often do you use digital slides?	Daily; Weekly; Several times per month; Rarely
Usage Context	In which contexts do you primarily use digital slides?	Teaching and assessment; Clinical diagnosis and case discussion; Not used
Perceived Impact on Assessment	How do you perceive the impact of digital slides on your assessment performance?	Decrease; No impact; Improvement
Functional and Application Value	Overall rating of digital slides’ functionality and application value (including core operations, auxiliary management, teaching application, and economic value)	1–5 points
User Evaluation	• Advantages of digital slides	Remote collaboration and digital management; Slide preservation and reuse
	• Limitations of digital slides	Dependence on technology and environment; Cost and learning curve
Acceptance	Teaching slides; Examination slides; Routine slide reading	Disagree; Neutral; Agree
Open-ended Question	What are the limitations of digital slides and areas for improvement?	Open-ended response

Frequency of Use indicates the regularity of participants’ engagement with digital slides

Usage Context identifies the primary settings in which digital slides are applied

Perceived Impact on Assessment evaluates participants’ perception of how digital slides influence performance outcomes

Functional and Application Value captures a comprehensive evaluation of digital slides, including core operations, auxiliary management, teaching applications, and economic value. Scores range from 1 (lowest) to 5 (highest)

User Evaluation explores both perceived advantages and limitations of digital slides

Acceptance assesses the degree of agreement with digitalization in different slide use contexts, measured on a three-point Likert scale: Disagree, Neutral, Agree

Open-ended Question allows respondents to elaborate on limitations and suggest areas for improvement

Statistical analysis

Data were analyzed using SPSS 27.0, with two-sided significance tests set at P < 0.05. Categorical variables were summarized as counts (n) and percentages (%) and compared using the chi-square test or Fisher’s exact test. Continuous variables were reported as mean ± standard deviation (Mean ± SD) and compared using independent-sample t-tests or one-way ANOVA. Hierarchical multiple linear regression was conducted to explore potential independent associations between assessment modality and scores, with assessment score as the dependent variable and residency year, educational background, and assessment modality as predictors. Because assessment modality was closely linked to assessment year, including both variables in the same model resulted in collinearity (VIF = 5.170). Therefore, the assessment year was not included in the final regression model. The potential for residual temporal confounding was acknowledged when interpreting the findings. Regression residuals were evaluated for normality (Shapiro–Wilk) and homoscedasticity (Breusch–Pagan test). Questionnaire reliability and validity were examined via Cronbach’s α and EFA, with factor scores subjected to normality and variance tests. Differences in digital slide acceptance across residency years were assessed using the Kruskal–Wallis H test.

Results

Retrospective assessment results

The baseline characteristics of the participants in the two assessment modalities are presented in Table 2, with multiple linear regression results summarized in Table 3. After adjustment for residency year and educational background, assessment modality showed no significant independent association with assessment scores (β = -0.015, P = 0.878). In contrast, residency year (β = 0.607, P < 0.001) and educational background (β = 0.218, P = 0.030) were significantly associated with assessment scores. These findings lend tentative preliminary reference for evaluating the feasibility of digital slide-based assessment within the context of the archived overall slide-reading assessment score. However, because component-level diagnostic subscores were unavailable, the results should not be interpreted as evidence regarding specific diagnostic performance domains.

Table 2.

Comparison of Assessment Scores and Baseline Characteristics of Resident Physicians under Different Assessment Modalities

Baseline Characteristics	Conventional Microscopy Group (n = 31)	Digital Slide Group (n = 33)
Assessment Score (Mean ± SD)	74.65 ± 13.45	73.80 ± 11.34
Residency Year (n)
Year 2	16	18
Year 3	15	15
Educational Level (n)
Bachelor’s degree or below	12	13
Master’s degree or above	19	20
Assessment Year (n)
2021–2022	31	0
2023–2024	0	33

Values are presented as mean ± standard deviation (SD) or number of participants (n)

Assessment score represents the examination performance of resident physicians

Residency year indicates the year of standardized residency training at the time of assessment (e.g., Year 2 refers to residents who were in their second year of training when they undertook the examination)

Educational level refers to the highest academic qualification attained by participants

Assessment year denotes the calendar year in which the examination was conducted. The conventional microscopy group included residents assessed in 2021–2022, whereas the digital slide group included residents assessed in 2023–2024

Table 3.

Hierarchical regression analysis of factors associated with assessment scores of resident physicians

Predictor	B	SE	β	P value	95% CI for B		VIF
					Lower	Upper
Model 1
Constant	28.684	6.988		< 0.001	14.711	42.657
Residency year	14.886	2.372	0.608	< 0.001	10.143	19.630	1.002
Educational level	5.453	2.426	0.218	0.028	0.601	10.305	1.002
Model 2 (+ assessment modality)
Constant	29.273	8.009		< 0.001	13.252	45.293
Residency year	14.876	2.392	0.607	< 0.001	10.090	19.661	1.003
Educational level	5.451	2.446	0.218	0.03	0.558	10.344	1.002
Assessment modality	-0.369	2.386	-0.015	0.878	-5.142	4.405	1.001

The dependent variable was the assessment score

Model 1 included residency year and educational level as predictors. Model 2 additionally included the assessment modality

B represents the unstandardized regression coefficient, SE the standard error, β the standardized regression coefficient, CI the confidence interval, and VIF the variance inflation factor used to assess multicollinearity among predictors

Assessment modality was coded as 1 = traditional microscopy and 2 = digital slides

Model comparison showed that the addition of assessment modality did not significantly improve the model fit (ΔR² = 0.000, ΔF(1, 60) = 0.024, P = 0.878). The Durbin–Watson statistic was 1.791. Residual diagnostics indicated that the assumptions of linear regression were satisfied, including normality of residuals (Shapiro–Wilk test: W = 0.985, P = 0.652) and homoscedasticity (Breusch–Pagan test: Model 1: F = 1.776, P = 0.178; Model 2: F = 2.500, P = 0.068)

Questionnaire survey results

General survey results

A total of 59 valid questionnaires were obtained in this study, including 12 first-year residents, 17 second-year residents, and 30 third-year residents. Table 4 summarizes the usage of digital slides. Overall, 66.10% of respondents reported using digital pathology slides, with teaching and assessment being the most frequent applications (approximately 86.44%). Regarding advantages, “facilitating remote collaboration and digital management” received full recognition from all participants (100%), and “advantages in slide protection and reuse” also received a high rating (84.75%). In terms of perceived limitations, resident physicians predominantly expressed concerns about the strong “technical and environmental dependence” of digital slides (91.53%), such as reliance on equipment and network, time-consuming scanning, and insufficient resolution, in line with findings from the open-ended questionnaire. Over half of the respondents (52.54%) reported concerns about cost and the learning curve, which may partly contribute to resistance to the routine adoption of digital slides.

Table 4.

Analysis of digital slide usage among resident physicians by residency year

Category / Item	Year 1 (n = 12) n(%)	Year 2 (n = 17) n(%)	Year 3 (n = 30) n(%)	χ²	P value
Frequency of use
At least several times per month	3(25.00)	13(76.47)	23(76.67)	11.357	0.003
Rarely	9(75.00)	4(23.53)	7(23.33)
Usage context
Teaching and assessment	6(50.00)	17(100.00)	28(93.33)	17.478	<0.001
Clinical diagnosis and case discussion	1(8.33)	8(47.06)	11(36.67)	4.917	0.083
Not used	6(50.00)	0	2(6.67)	17.478	<0.001
Perceived advantages
Remote collaboration and digital management	12(100.00)	17(100.00)	30(100.00)
Slide preservation and reuse	9(75.00)	13(76.47)	28(93.33)	3.494	0.130
Perceived limitations
Dependence on technology and environment	9(75.00)	16(94.12)	29(96.67)	5.395	0.082
Cost and learning curve	9(75.00)	8(47.06)	14(46.67)	3.047	0.247

Values are presented as n (%). For usage context and perceived advantages/limitations, multiple responses were allowed; therefore, percentages within a section may exceed 100%. Differences among residency years were analyzed using the chi-square test or Fisher–Freeman–Halton exact test, as appropriate

Frequency of use refers to how often residents reported using digital slides

Usage context describes the primary settings in which digital slides were applied

Advantages and limitations represent participants’ perceptions of the benefits and challenges associated with the use of digital slides

Reliability and validity analysis of the questionnaire scale

A between-group comparison of the two factors identified through exploratory factor analysis was conducted among resident physicians based on their year of training, with results shown in Table 5. The analysis found significant differences in the evaluation of digital pathology slides regarding the “diagnostic operation efficiency” dimension across different residency years (F = 4.614, P = 0.014). The average scores for this factor showed a decreasing trend as residency year increased (Year 1: 0.525 > Year 2: 0.256 > Year 3: -0.355). This may imply that junior residents rely more heavily on basic functions of digital pathology, such as ease of operation, image clarity, and diagnostic speed, consistent with the “novice” stage of Dreyfus’ skill acquisition model [4]. This pattern was referred to as the “novice tool dependence” effect in this study. Conversely, no significant differences were observed in the evaluation of the “management and collaboration efficacy” dimension among resident physicians at different training stages, implying that the perceived value of digital slides in teaching, resource management, and team collaboration remains stable regardless of residency year. This may tentatively suggest that resident physicians generally recognize the importance of digital pathology for collaboration and management.

Table 5.

Evaluation of user experience with digital slides among resident physicians

	Year 1 Mean ± SD	Year 2 Mean ± SD	Year 3 Mean ± SD	Test Statistic	P value
Management collaboration efficiency	-0.283 ± 0.987	0.192 ± 0.528	-0.004 ± 1.192	H = 3.913	0.141
Diagnostic operation efficiency	0.525 ± 0.869	0.256 ± 0.907	-0.355 ± 0.989	F = 4.614	0.014

Factor scores were derived from exploratory factor analysis; higher scores indicate stronger endorsement of the corresponding construct. Values are presented as mean ± SD. The Shapiro-Wilk test was used to assess normality of factor scores; Levene’s test was used to assess homogeneity of variance. Group differences across residency years were analyzed using one-way ANOVA when normality and homogeneity assumptions were met and using the Kruskal–Wallis test otherwise

Analysis of acceptance of digital slide applications

The Kruskal-Wallis H non-parametric test was used to analyze differences in the acceptance of digital slides among resident physicians at various training levels, with results shown in Table 6. The analysis found significant differences in the acceptance of the “digitalization of routine slide reading” among residents by year (H = 6.636, P = 0.036). Specifically, the mean rank of acceptance was lower for third-year residents (24.92) compared to first-year (33.83) and second-year residents (36.26). For “digitalization of examination slides”, no statistically significant difference was observed among residents of different years (H = 5.580, P = 0.061). However, a trend toward a difference was noted: second-year residents had the highest mean rank of acceptance (36.38), while third-year residents had the lowest (26.65). This trend may suggest that “mid-stage resident physicians have relatively stronger adaptability to the new assessment mode” in clinical practice training. No significant difference was observed in acceptance of the digitalization of teaching slides across residency years (H = 2.598, P = 0.273).

Table 6.

Acceptance of digital slide applications among resident physicians across different residency years

Application scenario	Year 1 n(%)	Year 2 n(%)	Year 3 n(%)	H statistic	P value	Mean rank (Year 1 / Year 2 / Year 3)
Digitalization of teaching slides	10(83.33)	17(100.00)	27(90.00)	2.598	0.273	27.67/32.50/29.52
Digitalization of examination slides	8(66.67)	16(94.12)	18(60.00)	5.580	0.061	29.33/36.38/26.65
Digitalization of routine slide reading	8(66.67)	12(70.59)	11(36.67)	6.636	0.036	33.83/36.26/24.92

Values are presented as n (%) for residents selecting “agree” for each application scenario

Inferential comparisons across residency years were performed using the full ordinal responses of the three-point scale (disagree/neutral/agree) with the Kruskal–Wallis test

Mean rank values indicate the relative ordering of acceptance across residency years, with higher mean ranks indicating greater acceptance

Open-ended responses

Residents generally reported that digital slides improved operational convenience, preservation, multi-user collaboration, and self-directed learning. They also facilitated communication with supervising physicians and enhanced diagnostic ability. However, digital slides lagged behind traditional glass slides in diagnostic speed and image clarity, with frequent mentions of system/network/operation issues (29 times), resolution/image clarity (24 times), and equipment/scanning (19 times). Traditional glass slides remained strongly favored because of their clear advantages, such as “stronger intuition and no dependence on equipment,” making it hard to change respondents’ long-standing diagnostic habits. Based on the subjective feedback from surveyed participants, most respondents considered digital slides suitable for teaching and assessments, but not yet a replacement for routine clinical diagnosis.

Discussion

With the rapid advancement of information technology and digital transformation across healthcare, digital solutions have gained increasing adoption within clinical medicine [5–8]. In this context, the concept of “construction of digital pathological discipline” has emerged [1, 9]. As a vital component of digital pathology departments, whole-slide imaging (WSI) has overcome time and space limitations, providing a critical technical foundation for integrating teaching and assessment into standardized residency training (SRT). This study employed a combination of objective and subjective measures to preliminarily assess the application of digital pathology, which offers tentative reference regarding its value in supporting residents’ daily work, teaching, and assessment, rather than definitive evidence to inform pathology residency training reform.

Application and practice of digital slides in daily work

As early as 2003, institutions such as the University of California, Los Angeles (UCLA) School of Medicine and the University of Iowa Carver College of Medicine pioneered the integration of “virtual microscope slides” into their pathology courses, predicting that digital pathology could eventually replace conventional microscopes and photomicrographs in educational settings [10, 11]. A 2019 study [12] in the United States reported that the agreement between digital slides and traditional glass slides on key diagnostic indicators reached 99.3%. This suggests that digital slides reliably reproduce critical diagnostic information and can serve as an efficient adjunct to conventional pathology workflows, consistent with other research findings [13, 14].

Advantages of digital slides in residency training and assessment

Standardized resource supply allows digital slides to effectively solve issues caused by uneven quality of traditional glass slides, improve assessment management efficiency, support process recording and remote evaluation [15, 16], and provide technical assistance for implementing formative assessment in pathology SRT. Additionally, their easy storage, shareability, and support for remote collaboration help improve the consistency and accessibility of SRT educational resources across multiple hospital campuses. This not only enhances teaching and learning efficiency but also addresses traditional problems such as loss and poor reusability of glass slides. Study data indicated that changing the assessment mode had no significant impact on scores (β = -0.015, P = 0.878), and over half of the resident physicians (about 54.24%) believed that the mode change did not affect their scores.

The consistency between residents’ perceptions and archived assessment outcomes provides preliminary support for the feasibility of incorporating digital slides into SRT assessment. However, these findings should not be interpreted as evidence of equivalence between digital slide–based assessment and conventional microscopy. Furthermore, this study identified the “novice tool dependence” effect in the functional evaluation of digital pathology technology among resident physicians at various training stages. Novice trainees rely more heavily on clear operational rules and basic tools initially, focusing mainly on fundamental functions. As they gain experience, their focus shifts from “whether the tool is easy to use” to “improving diagnostic skills with the tool.” The “novice tool dependence” effect may provide preliminary reference for customizing SRT curricula. Early training should emphasize core skills such as observing and annotating digital slides, while later stages should focus on analyzing difficult cases digitally, developing diagnostic thinking, and enhancing clinical decision-making. This approach helps residents progress toward the “competent” or even “proficient” level according to the Dreyfus model of skill acquisition.

Challenges in the comprehensive popularization of digital pathology in residency training

Despite the demonstrated advantages of digital pathology in teaching and assessment, its broad implementation faces several challenges [17, 18]. Our survey indicated that entrenched habits with traditional microscope slides appear to be a key contributing factor behind the low acceptance of digital slides among resident physicians. In addition, 91.53% of respondents reported that technical and environmental dependencies—including image resolution, network stability, and scanning time—pose significant obstacles. High initial investment and ongoing maintenance costs further limit widespread adoption.

Limitations

This study has several limitations. First, the comparison between conventional microscopy and digital slide assessment was conducted across different examination periods because digital assessment was introduced as part of an institutional educational reform. Therefore, residual temporal confounding cannot be fully excluded. Changes in curriculum structure, faculty involvement, assessment organization, case selection, or resident cohort characteristics may have influenced the observed assessment outcomes. Accordingly, the absence of a statistically significant difference in overall scores should be interpreted as preliminary evidence supporting feasibility, rather than definitive evidence of equivalence between the two assessment modalities. Second, the retrospective performance analysis relied on archived overall composite assessment scores, and detailed diagnostic component scores were unavailable. Therefore, the study could not determine whether digital slides influenced specific diagnostic domains, such as morphologic description, diagnostic accuracy, differential diagnosis, or diagnostic reasoning. Third, the sample size was relatively modest, with 64 assessment records and 59 valid questionnaire responses, which may have limited the statistical power to detect small differences between assessment modalities or among residency-year subgroups.

In addition, performance data were derived from a single training center, and survey responses were collected from SRT bases within one province. These sampling characteristics may limit transferability to other regions or institutions with different digital pathology infrastructure, faculty experience, and residency training systems. Furthermore, although the questionnaire supported acceptable internal consistency and exploratory factor structure, it was developed for the present study and has not undergone external psychometric validation. Therefore, the survey findings should be interpreted as exploratory. Finally, open-ended responses were summarized descriptively rather than analyzed using a formal qualitative framework. Despite these limitations, the present findings provide useful preliminary insights into the educational application of digital pathology in SRT and may help inform future multicenter investigations involving larger cohorts and more detailed educational outcome measures.

Future directions

Building upon these preliminary findings, several directions for future development and research warrant consideration. Our finding that residents demonstrated higher acceptance of digital slides for teaching and assessment than routine diagnostic reading aligns with the staged implementation trend widely reported across digital pathology practice, where digital platforms are commonly rolled out for education and consultation before routine diagnostic application. As summarized in a comprehensive WSI practical guideline, digitized slides build an indispensable data infrastructure to underpin the development of pathology artificial intelligence [19]. Combining our empirical outcomes with published domestic and international research [20–25], alongside emerging literature focusing on digital pathology-based AI development [26–28], future improvements to pathology SRT can prioritize the development of AI-assisted diagnostic and scoring systems and iterative refinement of existing training and assessment frameworks. Such targeted advancement is expected to reduce trainees’ adaptive resistance toward digital slide reading, elevate diagnostic and examination accuracy, and fully exploit the strengths of digital technology to optimize pathology residency education.

In summary, digital pathology, with its ability to overcome temporal and spatial limitations and to standardize teaching resources, demonstrates clear applicability to residency training and assessment. This study offers preliminary empirical evidence to support the digital reform of pathology education, although full integration into routine clinical diagnostics will require ongoing solutions to technical and cost-related challenges.

Authors’ contributions

Wangwang Liu: Conceptualization, Data curation, Formal analysis, Funding acquisition, Writing – original draft. Yi Gan: Investigation, Data collection, Validation. Qiuxia Gu: Methodology, Survey design, Formal analysis. Dan Qiao: Resources, Data curation, Project administration. Zhinong Jiang: Supervision, Methodology refinement. Wei Shi: Conceptualization, Supervision, Writing – review & editing, Corresponding author. All authors read and approved the final manuscript.

Funding

This research was supported by the Education Reform Project of Zhejiang University School of Medicine, 2025 (Grant No. jgyb2025068).

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Sir Run Run Shaw Hospital (protocol code 0213, approval date March 5, 2026). Participation in the survey was voluntary and anonymous. Completion of the questionnaire was considered as implied informed consent. The Institutional Review Board waived the requirement for written informed consent due to the anonymous nature of the survey and minimal risk to participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.