Evaluating a specialty-specific objective structured clinical examination and workplace-based assessment curriculum for radiation oncology postgraduate training: a prospective educational intervention study

Abstract

Background

Competence in radiation oncology is still commonly expected to emerge through apprenticeship, although essential tasks such as target delineation, treatment-plan review, and multidisciplinary reasoning are not reliably sampled by end-of-rotation examinations alone.

Methods

We conducted a single-center prospective educational intervention at a tertiary academic cancer center, using a historical-control comparison. 21 radiation oncology postgraduates completed the specialty-specific curriculum and were compared with 19 historical controls. The curriculum was built around 3 fixed disease modules: locally advanced nasopharyngeal carcinoma without induction chemotherapy, middle/lower thoracic esophageal squamous cell carcinoma managed with definitive or neoadjuvant radiotherapy, and FIGO 2018 IB3-IIIC1 cervical cancer treated with definitive chemoradiation. Each module combined contouring, structured plan review using FCB-CHOPS, MDT-style debriefing, and narrative feedback. Outcomes included OSCE, mini-CEX, multisource evaluation, FCB-CHOPS, and PHEEM.

Results

The intervention cohort had a higher OSCE total score than historical controls (79.4 ± 6.8 vs. 76.2 ± 7.0, P = 0.021). The contouring station differed slightly between groups (P = 0.049), whereas the plan-evaluation station did not (P = 0.100). Mini-CEX overall score (P = 0.006), FCB-CHOPS (P = 0.008), and PHEEM total score (P = 0.020) all increased after the intervention, while no significant change was observed in the 360-degree overall score or the PHEEM social-support domain.

Conclusions

This curriculum was associated with modest, uneven gains in selected domains of postgraduate radiation oncology training. Its main educational value may lie in making contouring and plan review more explicit, observable, and discussable within routine clinical teaching.

Background

Radiation oncology training depends on competence in several high-stakes tasks that are not well captured by written examinations alone, including target delineation, treatment-plan review, and multidisciplinary clinical reasoning. The 2022 Chinese residency-training standards and the companion teaching-activity guidance explicitly shifted postgraduate education toward structured, competence-oriented assessment and formally incorporated tools such as mini-clinical evaluation exercise (mini-CEX) and 360-degree evaluation into workplace teaching [1, 2]. That policy direction is consistent with Miller’s framework and later work on programmatic assessment, both of which emphasize that no single test adequately represents clinical competence [3–6]. A written examination may capture what a trainee knows; it cannot show whether the same trainee can contour a skull-base target safely, recognize a flawed registration, or explain to a physicist why a plan should be revised.

The weakness is particularly visible in treatment-plan review. In a recent national survey of radiation oncology trainees and faculty, nearly half of trainees reported reviewing fewer than half of their own patients’ plans, more than half felt that plan-evaluation training was inadequate, and a substantial proportion rarely reviewed plans directly with attending physicians [7]. Other contemporary studies have shown a similar gap between how faculty perceive plan education and how trainees experience it in daily practice [8]. This mismatch is educationally important because plan review is not a peripheral activity. It is the point at which anatomy, prescription intent, image fusion, target definition, organ-at-risk limits, and clinical priorities become operational. If that process remains tacit, trainees may appear clinically engaged while still learning the core task slowly and inconsistently.

Target delineation shows the same problem from another angle. Prospective contour-concordance studies have demonstrated measurable discrepancies between trainee contours and faculty reference contours, with substantial variation by disease site and by level of training [9]. Earlier prospective teaching work in head-and-neck contouring reached a similar conclusion: trainees improved after structured instruction, but improvement was neither uniform nor self-sustaining without deliberate practice and feedback [10]. In other words, radiation oncology trainees do not reliably “pick up” contouring skill simply by being present in a planning environment. What improves performance is repeated exposure to well-specified cases, comparison with accepted standards, and feedback that is specific enough to alter the next attempt.

For education research, the implication is straightforward: a single end-of-rotation score is unlikely to represent competence in this specialty. A combination of standardized assessment and workplace-based sampling is more defensible, particularly when the target skills are task-specific and high stakes.

We therefore developed and evaluated a specialty-specific curriculum that combined objective structured clinical examination (OSCE) with workplace-based assessment (WBA) across three fixed disease modules. Our objective was to determine whether a structured curriculum could make contouring and plan review more explicit and assessable within routine postgraduate training.

Methods

This was a single-center prospective educational intervention conducted in the Department of Radiation Oncology of a tertiary academic cancer center. The intervention cohort consisted of 21 postgraduates rotating through the department between December 2023 and December 2025. A historical-control cohort of 19 postgraduates trained during the past three academic years under the conventional model was used for comparison. The conventional model comprised attending-led case discussion, ad hoc contour review during routine planning, and an end-of-rotation oral examination, but did not include a fixed OSCE blueprint, a structured plan-review checklist, or scheduled narrative-feedback sessions. Concurrent randomization was not feasible within the departmental rotation schedule; therefore, a historical-control design was used, as in other recent radiation oncology education studies [11, 12].

Participants were eligible if they were postgraduate trainees in radiation oncology or closely related oncology tracks, had completed at least 2 weeks of department exposure, and consented to educational evaluation. Trainees were excluded if they rotated for fewer than 4 weeks, failed to complete baseline assessment, or had already undertaken a formal departmental contouring-and-plan-review curriculum.

Curriculum design

The curriculum was designed backward from core clinical tasks rather than lecture topics. Each disease module included one contouring task, one structured plan-review task, one multidisciplinary team (MDT)-style case presentation, and one narrative-feedback encounter.

The nasopharyngeal carcinoma module was restricted to locally advanced disease treated without induction chemotherapy. This decision was deliberate. Recent consensus work has shown that target delineation after induction chemotherapy involves distinct response-adapted logic, which can obscure basic contouring principles for junior trainees [13, 14]. The teaching package therefore used pre-radiotherapy computed tomography-magnetic resonance imaging registered datasets, disease summaries, and prompts aligned with contemporary primary-target guidance for nasopharyngeal carcinoma (NPC) [13, 15].

The esophageal module included middle/lower thoracic squamous cell carcinoma treated with either definitive chemoradiation or neoadjuvant radiotherapy. These cases were selected because they force trainees to address longitudinal tumor extent, elective nodal coverage, and cardiopulmonary trade-offs rather than relying on simple geometric expansion. Case materials and faculty scoring were aligned with expert contouring guidance for intensity-modulated radiation therapy in esophageal and gastroesophageal junction cancers, while cervical esophagus and junctional tumors were excluded to preserve anatomical consistency at this training stage [16].

The cervical module focused on International Federation of Gynecology and Obstetrics (FIGO) 2018 IB3-IIIC1 intact cervical cancer treated with definitive chemoradiation. This module was chosen because it requires integration of pelvic anatomy, organ motion, nodal target definition, and the logic of external-beam radiotherapy in relation to brachytherapy [17–19]. The curriculum stopped at IIIC1 rather than extending to para-aortic disease because the educational target was core pelvic competence, not extended-field planning.

Educational process

Before each module, trainees received a preparatory package consisting of one contouring atlas, one planning note, and a short summary of common pitfalls. They then completed a supervised contouring task and a separate plan-review task using an attending-approved plan and an intentionally flawed comparison plan. Structured review was based on Fusion, Contours, Beams, Coverage, Heterogeneity, Organs at Risk, Prescription, Summation (FCB-CHOPS).

After the technical tasks, trainees presented the case in MDT format to a faculty member and either a physicist or therapist. Discussion focused on target choice, trade-offs between target coverage and organ-at-risk sparing, and the rationale for either plan acceptance or revision. Faculty then provided narrative feedback directed at what had been missed, what had been prioritized incorrectly, and what should change at the next encounter. Prior work suggests that feedback is more effective when it is case-specific and actionable rather than limited to numerical ratings alone [20–23].

Each disease module was delivered over approximately 2 weeks and included 2–3 h of supervised contouring review, 2 h of structured plan-review discussion, one MDT-style case session, and one narrative-feedback session, yielding approximately 6–8 h of supervised faculty contact per module. The curriculum was delivered by four attending radiation oncologists, with participation from one medical physicist and one radiation therapist during MDT-style debriefing and plan-review teaching.

The overall study design and curriculum structure are summarized in Fig. 1.

Fig. 1.

Study design and curriculum schema. This figure summarizes the structure of the specialty-specific assessment curriculum developed for radiation oncology postgraduate training. Participants completed three fixed disease modules: locally advanced nasopharyngeal carcinoma without induction chemotherapy, middle/lower thoracic esophageal cancer treated with definitive or neoadjuvant radiotherapy, and FIGO 2018 IB3–IIIC1 cervical cancer treated with definitive chemoradiation. Within each module, trainees undertook contouring and plan-review tasks, followed by MDT debriefing and narrative feedback. Educational outcomes were evaluated using OSCE, mini-CEX, 360-degree evaluation, FCB-CHOPS, and PHEEM. Note: NPC: nasopharyngeal cancer; FIGO: International Federation of Gynecology and Obstetrics; OSCE: objective structured clinical examination; mini-CEX: mini-clinical evaluation exercise; FCB-CHOPS: fusion, contours, beams, coverage, heterogeneity, organs at risk, prescription, and dose summation; PHEEM: postgraduate hospital educational environment measure

Assessment strategy

The assessment framework deliberately combined standardized and workplace-based tools. OSCE functioned as the end-of-rotation summative assessment. Two 25-minute stations were used: a contouring station and a plan-evaluation station. Station blueprints were balanced across the three disease modules. Contouring performance was scored on target extent, anatomical boundary logic, organ-at-risk completeness, and verbal justification. Plan evaluation was scored on identification of target-coverage issues, detection of organ-at-risk violations, prioritization of competing objectives, and appropriateness of the final recommendation. Two faculty raters who were not involved in day-to-day supervision scored independently while blinded to cohort assignment.

WBA was sampled throughout the rotation. Mini-CEX was used during live case presentation or on-treatment decision discussions and focused on clinical reasoning, communication, prioritization, and organization. Multisource 360-degree evaluation was completed by attending physicians, medical physicists, therapists, and oncology nurses, with emphasis on teamwork, accountability, and professional behavior. In addition, a department-adapted FCB-CHOPS checklist was completed during each plan-review session to capture whether the trainee reviewed the plan systematically rather than selectively [24]. FCB-CHOPS included the following domains: image fusion, contour review, beam arrangement, target coverage, dose heterogeneity, organs at risk, prescription consistency, and dose summation. For each domain, trainees were scored on whether the issue was reviewed systematically and whether relevant problems were identified and prioritized appropriately during plan discussion. Because FCB-CHOPS was also used during teaching, it was treated as a process-oriented educational outcome reflecting the structure and completeness of plan review, rather than as a fully independent measure of transferable competence. This combination was intended to capture performance across settings rather than rely on a single end-of-rotation judgment, consistent with programmatic-assessment principles [6, 20, 25–27].

A faculty consensus approach was used to define station content, checklist domains, and scoring anchors before implementation. Because the study focused on comparative educational performance rather than pass/fail decision-making, OSCE scores were analyzed as continuous outcomes rather than being converted into a fixed pass standard.

Learning environment

The postgraduate hospital educational environment was measured using postgraduate hospital educational environment measure (PHEEM), which covers role autonomy, teaching quality, and social support [28]. PHEEM was selected because it has been widely used in postgraduate settings and was recently validated in a Chinese residency context [29]. A short departmental supplement addressing access to contouring opportunities, participation in plan-review discussion, and usefulness of feedback was added for local interpretation, but these items were analyzed descriptively rather than folded into the validated total score.

Outcomes and statistical analysis

The primary outcome was end-of-rotation OSCE total score. Secondary outcomes were contouring-station score, plan-evaluation-station score, mini-CEX, 360-degree overall score, FCB-CHOPS score, and PHEEM total and subdomain scores. OSCE total score was prespecified as the primary endpoint. All secondary analyses were exploratory and were intended to describe the multidimensional educational profile of the curriculum rather than to support definitive claims for each individual secondary outcome. Inter-rater reliability for OSCE was assessed with the intraclass correlation coefficient. Internal consistency of questionnaire-based instruments was assessed with Cronbach’s α. Continuous variables were compared with independent or paired t tests, depending on the contrast of interest, and categorical variables were analyzed with the chi-square or Fisher exact test. Two-sided P values below 0.05 were considered statistically significant. Continuous outcomes are reported as mean±standard deviation, together with mean differences or mean changes, 95% confidence intervals, and effect sizes where appropriate.

Results

Participant characteristics

The intervention cohort included 21 postgraduates and the historical-control cohort included 19. Baseline characteristics were comparable. Mean age was 26.1 ± 1.4 years in the intervention cohort and 27.1 ± 1.6 years in controls; women accounted for 71.4% and 73.7%, respectively. Median previous radiation-oncology exposure was 1 month in both cohorts, and baseline written oncology-knowledge scores were similar (78.4 ± 5.9 vs. 77.6 ± 6.3, P = 0.49).

OSCE outcomes

The intervention cohort achieved a higher OSCE total score than the historical-control cohort (79.4 ± 6.8 vs. 76.2 ± 7.0, P = 0.021). Improvement was observed in the contouring station (39.1 ± 4.1 vs. 37.4 ± 4.5, P = 0.049), whereas the difference in the plan-evaluation station did not reach statistical significance (39.3 ± 3.9 vs. 38.0 ± 4.0, P = 0.100). At the module level, the cervical cancer case showed a significant between-group difference (78.5 ± 6.3 vs. 75.8 ± 6.6, P = 0.037), whereas the NPC and esophageal modules showed only modest, non-significant trends (NPC: 78.1 ± 7.0 vs. 75.4 ± 7.4, P = 0.061; esophageal: 77.9 ± 6.5 vs. 76.0 ± 6.8, P = 0.153). NPC cases showed the widest score dispersion at baseline. In the esophageal module, the most frequent difficulty remained underestimation of longitudinal target extent and related cardiopulmonary trade-offs. The detailed data are presented in Table 1. A cross-domain visual summary of the main outcome measures is shown in Fig. 2.

Table 1.

OSCE outcomes

Outcome	Historical control (n = 19)	Intervention (n = 21)	Mean difference (95% CI)	Effect size (Cohen’s d)	P value
OSCE total score (0-100)	76.2 ± 7.0	79.4 ± 6.8	3.2 (0.49 ~ 5.91)	0.46	0.021
Contouring station (0–50)	37.4 ± 4.5	39.1 ± 4.1	1.7 (0.01 ~ 3.39)	0.40	0.049
Plan evaluation station (0–50)	38.0 ± 4.0	39.3 ± 3.9	1.3 (-0.25 ~ 2.85)	0.33	0.100
NPC case score (0-100)	75.4 ± 7.4	78.1 ± 7.0	2.7 (-0.13 ~ 5.53)	0.38	0.061
Esophageal case score (0-100)	76.0 ± 6.8	77.9 ± 6.5	1.9 (-0.72 ~ 4.52)	0.29	0.153
Cervical case score (0-100)	75.8 ± 6.6	78.5 ± 6.3	2.7 (0.16 ~ 5.24)	0.42	0.037

Data are presented as mean ± SD unless otherwise indicated. Mean difference was calculated as intervention minus historical control

OSCE objective structured clinical examination, NPC nasopharyngeal carcinoma

Fig. 2.

Cross-domain comparison of baseline and post-curriculum outcomes. The figure summarizes mean scores with 95% confidence intervals for the main study outcomes before and after implementation of the specialty-specific curriculum. The displayed outcomes include OSCE total score, plan-evaluation OSCE score, mini-CEX, FCB-CHOPS checklist score, 360-degree evaluation, and PHEEM total score. Blue bars represent baseline scores and orange bars represent post-curriculum scores. ^* indicate statistically significant between-timepoint differences (P < 0.05). Overall, the figure shows clearer improvement in structured and directly observed domains than in broader multisource or learning-environment measures. Note: OSCE, objective structured clinical examination; mini-CEX, mini-clinical evaluation exercise; FCB-CHOPS, fusion, contours, beams, coverage, heterogeneity, organs at risk, prescription, and dose summation; PHEEM, postgraduate hospital educational environment measure

Workplace-based assessment

The mini-CEX overall score improved from 6.4 ± 0.8 at baseline to 6.7 ± 0.8 at the end of rotation (P = 0.006). The largest gain was seen in clinical reasoning (6.1 ± 0.8 vs. 6.9 ± 0.7, P < 0.001), followed by communication (6.5 ± 0.7 vs. 6.8 ± 0.7, P = 0.002). The 360-degree overall score increased numerically from 4.03 ± 0.33 vs. 4.11 ± 0.31 but did not reach statistical significance (P = 0.063), while professional accountability showed no clear change over the rotation period (4.11 ± 0.31 vs. 4.17 ± 0.29, P = 0.135). FCB-CHOPS score improved from 21.9 ± 3.7 vs. 23.2 ± 3.5 (P = 0.008), suggesting a more structured approach to plan review. The detailed data are presented in Table 2.

Table 2.

WBA outcomes

Outcome	Baseline	End of rotation	Mean change (95% CI)	Effect size	P value
mini-CEX overall (1–9)	6.4 ± 0.8	6.7 ± 0.8	0.30 (0.09 ~ 0.51)	0.37	0.006
Clinical reasoning (1–9)	6.1 ± 0.8	6.9 ± 0.7	0.80 (0.35 ~ 1.25)	1.06	< 0.001
Communication (1–9)	6.5 ± 0.7	6.8 ± 0.7	0.30 (0.12 ~ 0.48)	0.43	0.002
Organization/efficiency (1–9)	6.3 ± 0.8	6.6 ± 0.8	0.30 (0.09 ~ 0.51)	0.37	0.006
360-degree overall (1–5)	4.03 ± 0.33	4.11 ± 0.31	0.08 (-0.01 ~ 0.17)	0.25	0.063
Team communication (1–5)	3.97 ± 0.36	4.12 ± 0.34	0.15 (0.06 ~ 0.24)	0.43	0.002
Professional accountability (1–5)	4.11 ± 0.31	4.17 ± 0.29	0.06 (-0.02 ~ 0.14)	0.20	0.135
FCB-CHOPS (0–32)	21.9 ± 3.7	23.2 ± 3.5	1.30 (0.35 ~ 2.25)	0.36	0.008

Data are presented as mean ± SD unless otherwise indicated. Mean change was calculated as end of rotation minus baseline

WBA workplace-based assessment, mini-CEX mini-clinical evaluation exercise, FCB-CHOPS fusion, contours, beams, coverage, heterogeneity, organs at risk, prescription, and dose summation

Correlation analysis

End-of-rotation OSCE total score showed a moderate correlation with FCB-CHOPS score (r = 0.61, P < 0.001) and a somewhat weaker but still meaningful correlation with mini-CEX (r = 0.48, P < 0.001). These findings indicate that OSCE and workplace-based tools captured overlapping but not identical aspects of competence. Higher OSCE scores were generally accompanied by more structured plan review in the clinical setting.

Learning environment and reliability

PHEEM total score increased from 111.2 ± 13.1 vs. 115.3 ± 12.8 (P = 0.020). Role autonomy and teaching improved modestly (34.1 ± 4.9 vs. 35.5 ± 4.8, P = 0.033; 44.0 ± 5.4 vs. 45.8 ± 5.2, P = 0.013, respectively), whereas the social-support domain showed no clear change over time (33.1 ± 4.3 vs. 33.9 ± 4.1, P = 0.154). On the departmental supplement, 76.9% of trainees agreed that the curriculum created more meaningful opportunities to discuss plan trade-offs with faculty, and 73.1% agreed that feedback became more specific and more useful for the next case.

Inter-rater reliability for OSCE total score was high (intraclass correlation coefficient (ICC) = 0.87). Cronbach’s α was 0.91 for PHEEM in the present sample, 0.88 for the 360-degree instrument, and 0.86 for the departmental supplement, supporting acceptable internal consistency for an educational intervention study. The detailed data are presented in Table 3. Module-specific outcome patterns across the three fixed disease modules are illustrated in Fig. 3.

Table 3.

Learning environment outcomes

Outcome	Baseline	End of rotation	Mean change (95% CI)	Effect size	P value
PHEEM total (0-160)	111.2 ± 13.1	115.3 ± 12.8	4.1 (0.67 ~ 7.53)	0.32	0.020
Role autonomy	34.1 ± 4.9	35.5 ± 4.8	1.4 (0.12 ~ 2.68)	0.29	0.033
Teaching	44.0 ± 5.4	45.8 ± 5.2	1.8 (0.40 ~ 3.20)	0.34	0.013
Social support	33.1 ± 4.3	33.9 ± 4.1	0.8 (-0.33 ~ 1.93)	0.19	0.154

Data are presented as mean ± SD unless otherwise indicated. Mean change was calculated as end of rotation minus baseline

PHEEM postgraduate hospital educational environment measure

Fig. 3.

Module-specific comparison of pre-curriculum and post-curriculum outcomes across the three fixed disease modules. The figure shows mean scores with 95% confidence intervals for key outcome measures across the three predefined disease modules: locally advanced nasopharyngeal carcinoma without induction chemotherapy, middle/lower thoracic esophageal cancer treated with definitive or neoadjuvant radiotherapy, and FIGO 2018 IB3–IIIC1 cervical cancer treated with definitive chemoradiation. The displayed domains include OSCE total score, plan-evaluation OSCE score, mini-CEX, FCB-CHOPS checklist score, 360-degree evaluation, and PHEEM total score. Blue bars represent pre-curriculum scores and orange bars represent post-curriculum scores. ^* indicate statistically significant pre/post comparisons within the corresponding module (P < 0.05). The figure illustrates a non-uniform pattern of change across modules, supporting the interpretation that the fixed-module curriculum sampled different dimensions of competence rather than producing equivalent gains across all disease sites. Note: FIGO, International Federation of Gynecology and Obstetrics; OSCE, objective structured clinical examination; mini-CEX, mini-clinical evaluation exercise; FCB-CHOPS, fusion, contours, beams, coverage, heterogeneity, organs at risk, prescription, and dose summation; PHEEM, postgraduate hospital educational environment measure

Discussion

This study was based on a practical premise: in radiation oncology, contouring and plan review should be taught and assessed explicitly rather than assumed to develop through routine clinical exposure alone. The findings showed improvement in some domains but not in others. Overall, the observed effect sizes were small to moderate, suggesting a modest educational effect concentrated in selected competencies rather than a broad transformation across all measured domains. These results are more consistent with a targeted improvement in directly taught tasks than with a uniform shift in overall clinical performance [5, 6].

This mixed pattern is plausible in postgraduate training. Skills that are directly taught and directly observed, such as case-specific reasoning and structured plan review, may change earlier than broader interpersonal or reputational dimensions captured by multisource evaluation. In this study, OSCE total score, mini-CEX, and FCB-CHOPS showed clearer movement than 360-degree outcomes, which suggests that the different tools were sampling related but non-identical aspects of competence. The statistically significant findings should therefore be interpreted as evidence of a measurable educational signal, not as proof of large competence gains across all domains [5, 6, 30, 31].

The disease modules likely contributed to this mixed pattern. Locally advanced nasopharyngeal carcinoma, middle/lower thoracic esophageal cancer, and intact cervical cancer were selected because they expose different weaknesses in trainee reasoning, but they are not equally amenable to short-term teaching. NPC contouring is anatomically intricate and benefits from repeated exposure; esophageal planning requires comfort with longitudinal target extent and cardiopulmonary trade-offs; cervical cancer more readily lends itself to explicit teaching of pelvic target logic once imaging and nodal patterns are made concrete [13–18]. The fixed-module approach may therefore be most useful as a blueprint for sampling different dimensions of competence rather than as a mechanism for producing uniform gains across all disease sites within a limited rotation.

The plan-review findings also deserve cautious interpretation. Contemporary survey data suggest that treatment-plan education remains one of the least standardized components of radiation oncology training [7, 8]. In the present study, the plan-evaluation OSCE station did not show a strong independent signal, but FCB-CHOPS scores improved. This should not be interpreted too broadly. Because FCB-CHOPS functioned both as a teaching scaffold and as an assessment framework, improvement in this score may reflect not only better plan-review reasoning but also increasing familiarity with a structured sequence of review. We therefore regard FCB-CHOPS primarily as a process-oriented outcome indicating more disciplined and reproducible plan-review behavior, rather than as a stand-alone measure of generalized competence [23].

The learning-environment findings should likewise be interpreted with restraint. PHEEM total score improved, with modest gains in role autonomy and teaching, whereas the social-support domain showed no clear change over time. This is a credible pattern in educational intervention studies. A structured curriculum may clarify expectations, improve perceived teaching quality, and increase trainee ownership of tasks without necessarily altering the wider social environment of a department over a short time frame. For postgraduate radiation oncology training, this distinction matters: improvements in instructional structure and supervised autonomy may precede any broader change in perceived departmental support [19, 21, 22, 28, 29]. In addition, because PHEEM was assessed within the same intervention cohort before and after curriculum implementation, the observed change cannot fully distinguish a genuine shift in the broader educational environment from increased satisfaction with the curriculum itself.

Several limitations should be acknowledged. First, this was a single-center study with a historical-control design, so cohort differences, faculty effects, and secular changes in supervision cannot be fully excluded. Second, the assessment strategy relied partly on direct observation, which remains susceptible to generosity bias, halo effects, and variation in scoring thresholds despite rater preparation [20, 32]. Third, some outcomes—particularly 360-degree ratings—may have been insufficiently sensitive to detect small short-term changes over a single rotation. In addition, because the OSCE stations were aligned with the same disease modules used in teaching, some degree of case familiarity is possible; the OSCE findings should therefore be interpreted as evidence of structured performance within sampled domains rather than unrestricted transfer across all disease settings. Finally, secondary outcomes were analyzed on an exploratory basis without formal adjustment for multiple comparisons, which increases the possibility of false-positive findings beyond the prespecified primary endpoint.

Despite these limitations, the study offers a potentially transferable framework for specialties in which competence is distributed across structured technical tasks and multidisciplinary workplace performance. Radiation oncology increasingly incorporates AI-assisted contouring, adaptive workflows, and faster planning systems, but these developments do not reduce the need for trainees to inspect contours critically, question plans systematically, and articulate trade-offs clearly [12, 33]. The value of the present work lies less in the claim of a universally generalizable effect size than in the curriculum architecture itself: a specialty-specific framework that makes contouring, plan review, and feedback more explicit and assessable within routine training. Broader validation, ideally across multiple institutions, will be necessary before stronger conclusions about effectiveness can be drawn.

Conclusions

This specialty-specific curriculum was feasible to implement and was associated with modest gains in selected domains of radiation oncology postgraduate training. Its main educational value may lie in making contouring and plan review more explicit, observable, and discussable within routine clinical teaching. Broader effectiveness should be confirmed in larger, preferably multicenter, studies.

Abbreviations

FCB-CHOPS

Fusion, contours, beams, coverage, heterogeneity, organs at risk, prescription, and dose summation

FIGO

International Federation of Gynecology and Obstetrics

ICC

Intraclass correlation coefficient

MDT

Multidisciplinary team

mini-CEX

Mini-clinical evaluation exercise

NPC

Nasopharyngeal carcinoma

OSCE

Objective structured clinical examination

PHEEM

Postgraduate hospital educational environment measure

WBA

Workplace-based assessment

Authors' contributions

CYM and JYZ conceived the study and designed the research framework. CYM, JC, YF, LL, and JWZ developed the specialty-specific curriculum and assessment instruments. JC, YF, LL, and JWZ participated in data collection and data management. CYM and JC performed the data analysis and interpretation. CYM wrote the manuscript. JYZ critically revised the manuscript and provided overall supervision. All authors read and approved the final manuscript.

Funding

This research was supported by grants from the National Natural Science Foundation of China (81602792); General Project of Jiangsu Provincial Health Commission (M2024004); Jiangsu Provincial Basic Science (Natural Science) Research General Project (23KJB310023); Jiangsu Provincial Maternal and Child Health Research Project (F202210); State Key Laboratory of Radiation Medicine and Radiation Protection Funding Project (GZK1202101); Youth Characteristic Technology Project of Clinical Diagnosis and Treatment Technology Innovation Project of the First Affiliated Hospital of Soochow University (2100201).

Data availability

The datasets analyzed in this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Soochow University. All procedures were conducted in accordance with the Declaration of Helsinki and relevant institutional regulations. Written informed consent to participate was obtained from all postgraduate trainees prior to enrollment. Participation in the study was voluntary and had no impact on trainees’ routine rotation requirements, academic standing, or formal summative evaluation. All data were de-identified before analysis.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.