Living and post‐mortem CT scans in the gross anatomy lab: A study investigating differences in first‐year medical students' exam performance and perceptions

Basic competency in radiological imaging is essential for physicians to identify and manage diseases. An optimal place in which to include imaging in the medical curriculum is during anatomy as students can correlate the 3D anatomy from their body donors with the 2D cross‐sectional anatomy. The goal of this project was to enhance first‐year medical students' knowledge of cross‐sectional imaging in the gross anatomy lab and to investigate whether there are benefits to learning cross sectional imaging via scans from body donors versus living individuals. Student participant performance was evaluated on laboratory practical examinations, CT image questions and spatial anatomical knowledge in the thorax and abdomen sections of gross anatomy. Students learned the cross‐sectional imaging during dissections where they accessed the images relevant to their study on Pacsbin, a web‐based Digital Imaging and Communication in Medicine viewer, via iPads. Results showed no statistically significant differences in practical examination scores, spatial anatomical knowledge, or identification of anatomical structures on CT image questions between participants who learned from images on body donors versus living individuals. In a questionnaire given at the end of the course, participants cited that the CT images improved their anatomical and imaging knowledge and that they felt better prepared to use imaging software and interpret diagnostic imaging results upon entering clerkships. While there were no differences in academic performance between the groups, positive outcomes regarding student perceptions of anatomical and imaging knowledge and preparedness for use of imaging software were identified in this study.


INTRODUC TI ON
Competency in radiology is an essential skill for graduating physicians as diagnostic imaging is crucial for physicians to identify and manage diseases. 1Unfortunately, student instruction in imaging remains insufficient in many medical schools as a result of numerous issues, including but not limited to, lack of time in the curriculum and faculty limitations of time and resources. 2A recent study demonstrated that imaging training was severely underrepresented in the curriculum of many Scottish medical schools, representing only 0.3% of the total hours devoted to instruction. 3 the United States (US), radiology is not a required clerkship and according to a recent report by Lee et al. 4 only 20% of US medical schools require radiology rotations.Similar to the national trend, medical students at the Larner College of Medicine (LCOM) at the University of Vermont (UVM) do not have a required radiology clerkship but can participate in radiology elective courses.
With respect to students' perspectives on training in imaging, a recent survey of 472 medical students representing 31 US medical schools reported that the majority of student respondents desired more exposure to imaging during their undergraduate medical training and that they perceived less confidence interpreting common imaging findings of the thorax. 5We obtained similar outcomes after surveying senior medical students at UVM.
][8][9][10][11][12][13][14] For example, Branstetter et al. 6 demonstrated that medical student exposure to radiology in the preclinical years of medical school increased student interest in pursuing radiology as a career.Another more recent study found that the introduction of imaging in the medical curriculum supported medical students' ability to identify and describe anatomical structures in images and improved students' ability to differentiate abnormal anatomy from normal anatomy. 13Furthermore, the introduction of imaging early in medical student training enhanced medical students' confidence in the use of standard DICOM viewers and improved students' perceptions on their abilities to identify major anatomical structures in stacks of computed tomography (CT) images. 14 important time frame in which students can begin to interpret anatomical structures in cross-sectional images is during dissection, when students can better correlate anatomical structures between 2D images and the 3D body.In particular, several reports have documented the impact of incorporating post-mortem cadaveric CT scans to aid students' anatomical knowledge.For example, Lufler et al. 9 found that first-year medical students who studied from cadaveric CT scans scored higher on anatomy practical examinations, performed better on questions associated with spatial anatomy, and received higher overall course grades than students who did not use CT scans.A more recent study also concluded that incorporating cadaveric CT scans into the general anatomy curriculum significantly improved student test scores in anatomy, and that the intervention was greatly appreciated by students. 12A separate study found that while the introduction of post-mortem CT scans of cadavers did not improve medical students' academic performance in radiology-based questions or anatomy in general, students generally regarded the intervention positively as it pertained to early exposure to CT imaging and clinical medicine through radiology. 11her groups have used cadaveric imaging to create integrated learning modules.These modules have been designed to enhance medical students' anatomical and radiological knowledge as well as appreciation for the anatomical relevance of clinical cases. 7,10In one example, Bohl et al. 7 used scans of post-mortem pre-embalmed cadavers to create clinically relevant anatomical cases for first-year medical students.Using results from student surveys, the authors found that their cadaver CT cases increased students' confidence for identifying anatomical structures on images, increased their confidence for identifying pathologies, and reduced their anxiety surrounding imaging. 7Similarly, Jacobson et al. 8 scanned several cadavers to create virtual patient cases based on imaging findings and data associated with the cadavers' medical histories.These cases were then used by first-year medical students to integrate their basic and clinical sciences knowledge. 8In another study, McBain et al. 10  The authors reported that students who viewed pathological findings from diagnostic images of their cadaver as opposed to students who viewed pathological findings from different cadavers exhibited a more humanistic approach to their cadaver and perceived that this form of imaging enhanced their anatomical understanding. 10 this study, the primary goal was to improve the imaging education that first-year medical students receive during their anatomical training at LCOM.Prior to the onset of the study, a needs assessment was conducted on 3rd and 4th year medical students at LCOM.Results from this questionnaire served as a basis for conducting the study as it illuminated student perceived gaps in imaging knowledge.In order to determine which imaging source best supported students' knowledge of cross-sectional anatomy, the authors incorporated both post-mortem cadaveric scans as well as living individual scans.This enabled the authors to address whether there is an academic advantage that first-year medical students gain from learning anatomy from CT scans on their body donor, thereby integrating the 3D human anatomy directly with the 2D scans, as opposed to learning the anatomy from CT scans of body donors they are not dissecting or from living individuals.This conceptual framework 15 addressed a knowledge gap, as this study is the first, to the knowledge of the authors, to assess whether cadaveric scans are superior to scans from living individuals toward improving student learning outcomes in anatomy and anatomical imaging.To support this curricular endeavor, we utilized the expertise of both radiology residents and senior medical students who guided students in imaging orientation and identification of anatomical structures.A secondary objective was to analyze data from post-intervention questionnaires to evaluate participant perceptions about the impact of the intervention on anatomical and imaging knowledge.Considering this was an exploratory study, we assessed performance on anatomy practical examinations and imaging related questions between students who learned the anatomy from CT scans on their body donor, a different body donor, or from a living individual.We anticipated no differences in group performance as CT images represented a small proportion (6%) of questions tested on the practical examinations and questions were on major anatomical structures (e.g.lungs, ventricles, atria, etc), for which the difference in resolution between body donors and living CT scans is not critical for identification.We also assessed participant perspectives and satisfaction about the intervention and anticipated that responses would vary depending on the source of the scan students used.While the information in this report reflects the results from study participants, the intervention was given to all first-year medical students to help inform future curricular practices around imaging education and to provide an equitable learning experience for all students.

Anatomy course details
The anatomy curriculum at LCOM is part of a larger 18-week course:

Educational objectives
The educational goals for the learner are listed below: 1. Increase early exposure to cross sectional imaging.
2. Improve knowledge of anatomy both generally and on crosssectional imaging.
3. Improve confidence using imaging software.

Study setting
This study was carried out in the classroom and in the gross anatomy laboratory during Blocks 4 (Thorax) and 5 (Abdomen/Pelvis).In the gross anatomy laboratory, students had access to scanned images using iPads (described in more detail below) and students with an internet connection were able to view scanned images on their personal devices outside of the laboratory.

Cadaver scanning and scans from a living patient
Post-mortem CT scans were conducted on 6 cadavers.Cadavers were embalmed prior to scanning with use of a commercially available embalming fluid: Maryland Embalming Fluid (Chemisphere Corporation, St Louis, MO).All cadavers were embalmed by a UVM employee trained and licensed in procedures for transporting and embalming the deceased.Cadavers were scanned using a Philips iCt 128 CT scanner, available at the Diagnostic Imaging Center at UVM. Images were rendered in 5 mm slices.Cadavers were transported from the Given Medical Building to the Diagnostic Imaging Center at the UVM hospital using a van specialized for transport of the deceased.Cadavers were covered at all times during transport.
Images of the chest, abdomen, and pelvis were acquired for each of the 6 cadavers and were obtained in the axial, coronal, and sagittal planes.Images from a healthy, living individual were obtained from de-identified UVM patient data.The patient data came from a noncontrast study.

Software for viewing CT images
Images were uploaded and viewed on Pacsbin, a web-based radiology teaching platform (Orion Medical Technologies; Eatontown, NJ).The Pacsbin DICOM viewer is free of charge and requires that users set up an account with a password.For the present study, body donors or living individual scans were uploaded on this platform as de-identified data to be in full compliance with patient privacy and protected health information standards.Depending on the group assignment, student participants were given access to the scans using a login and password specific for their study (e.g.anatomytable 1).Participants could access these scans in the anatomy laboratory on iPads or outside of the laboratory on their personal devices.

Participants and study groups
All first-year medical students enrolled in FoCS at LCOM in 2022 were informed of the imaging project via an in-class announcement and through two email communications.These communications were given in Blocks 3 (Lower Limb) and 4 (Thorax).Participation in the study was voluntary and students could opt out at any time during the course.There were no students who withdrew from the study.In total, 57 students provided written consent to participate in the project.The demographics included: 24 self-identified female, 19 self-identified male, 2 nonbinary, and 2 unspecified.Participants were placed in one of three groups based on the anatomy dissection table they were working at: CT Group 1: participants who learned the cross-sectional anatomy from images of body donors they were dissecting, CT Group 2: participants who learned the cross-sectional anatomy from images of body donors they were not dissecting and CT Group 3: participants who learned the cross-sectional anatomy from non-contrast images of a healthy living individual.Since this study was initiated toward the middle of FoCS, student dissection groups (A, B, C) and table numbers had already been established.
As such, there was an unequal distribution of participants per study group with 26 participants in CT Group 1, 12 participants in CT Group 2, and 18 participants in CT Group 3. The majority of participants signed up to participate at the beginning of the study; however, students had the opportunity to sign up at any point during the duration of the study.

Participant expectations
Participants were encouraged to label assigned anatomical structures (File S1) outside of the anatomy laboratory with faculty and senior medical students providing additional support.Here, a faculty member at LCOM prepared a short 10-min recording intended to guide students in the use of tools, including the annotation feature, on Pacsbin.In addition, faculty and senior medical students assisting with the project could utilize the students' assigned login and password to access their annotations and provide feedback to the student or group if requested.Although the authors originally anticipated that labeling activities could be done in the anatomy laboratory during dissection times, faculty and senior medical students were nonetheless able to engage with participants remotely and during optional learning sessions.

Instructional design and educational support
For each laboratory period in Blocks 4 (Thorax) and 5 (Abdomen/ Pelvis), participants were given a list of structures to identify on axial, coronal, and sagittal planes of CT images (File S1).Participants were also given a Pacsbin login and password corresponding to their assigned images.In preparation for the laboratories incorporating the CT imaging, participants were given three 10-min recorded modules introducing them to the features of Pacsbin (windowing, annotation, etc.) as well the process of identifying anatomical structures in the Thorax and Abdomen.During each 3-h laboratory session, 2-4 radiology residents were available to assist students with orientation and identification of anatomical structures on CT images.Outside of the laboratory, participants were encouraged to use the annotation feature on Pacsbin to label anatomical structures in the axial, coronal, and sagittal planes.In addition, senior medical students in their 3rd and 4th years of training held three optional review sessions for students.These sessions were intended to give students the opportunity to ask questions as well as to engage them in the learning process.Scrollable quizzes were also developed by the senior medical students participating in this project.These quizzes were prepared in Pacsbin and included content that students were responsible for identifying in Blocks 4 (Thorax) and 5 (Abdomen).
The quizzes were specific to each study group.Finally, e-anatomy (IMAIOS, New York, NY), an atlas of anatomy which includes cross sectional and MRI imaging, was available on all iPads in the anatomy lab as an additional resource to supplement learning.

Practical examinations
Participants were assessed using anatomy practical examinations.
Specifically, participants were evaluated on their ability to recognize tagged anatomical structures and answer the follow-up questions.Participants were also evaluated on their spatial anatomical knowledge (e.g.what structure was reflected in order to visualize the tagged structure or what structure is found immediately deep to the tagged structure?).Two spatial anatomical questions in the Thorax practical examination and 3 spatial anatomical questions in the Abdomen/Pelvis were included in each practical examination.Finally, participants were evaluated on their ability to identify CT-based questions.For each practical in Blocks 4 (Thorax) and 5 (Abdomen/Pelvis), two CT image questions were asked.Questions were made in Pacsbin and presented to students on the iPads.
Here, students could interact with the iPad during the exam, selecting links to the labeled structure in the axial, coronal, and sagittal planes (Figure 1).Students were also given a follow-up part B question which asked more detailed information about the CT structure (Figure 1).Students were encouraged to use the scrolling feature associated with the Pacsbin to visualize the progression of the labeled structure.In the Abdomen/Pelvis and Head and Neck practical examinations, one CT bonus question was included which tested prior knowledge on Thorax and Abdomen anatomy, respectively.Similar for all questions in the practical, students were given 70 s to answer parts A and B for CT-related questions.Each practical examination was graded out of 100% and the percentage of questions answered correctly was used in data analyses.Questions that used CT images were evaluated separately as well as combined in the total practical examination grade.

Questionnaires
All questionnaires associated with the study were administered using Qualtrics XM version 2020, (Qualtrics, Provo, UT), a webbased survey tool which is available free of cost to faculty and students at UVM.
A preintervention questionnaire was given to participants in this study.This 7-item questionnaire was designed to assess participants' prior knowledge and experience with anatomy and imaging prior to matriculation in addition to their major/minor in college (File S2).
At the conclusion of the course, participants were given an 18-item post-intervention questionnaire designed to assess participants' perceptions on whether the images improved their knowledge and understanding of gross anatomy and cross-sectional anatomy and their confidence using the Pacsbin platform amongst other items (File S3).Participants were queried on a 4-or 5-point Likert scale and in some cases, questions were presented as yes/no/ neutral (File S3).

Data collection
Written consent forms were obtained from participants.
Participant names were numerically coded for analysis of assessment data including practical examination scores and performance on CT and anatomical spatial knowledge questions.This data was collected for Blocks 4 (Thorax) and 5 (Abdomen/Pelvis) of FoCS.
Questionnaire data was obtained from Qualtrics XM and exported as an excel file (Microsoft, Redmond, WA).Questionnaire data reported for the post-intervention questionnaire was obtained anonymously and for the preintervention questionnaire, participants names were collected and data obtained was coded as previously described.

F I G U R E 1
Example CT image question from a body donor.The image was presented in the thorax practical on Pacsbin via an iPad.Students could select the plane in which they wished to visualize the labeled structure.In this image, the label is shown in the sagittal and coronal planes.A secondary (part B) question was given for each identification (part A) question.Cohen's d was determined using the formula below: M 1 is the mean practical performance for participants in one group, and M 2 is the mean of the practical performance for participants in the second group.Since there were three groups, the authors assessed Cohen's d between Groups 1 and 2, 1 and 3, and 2 and 3.This was done for anatomical practical examination performance in the thorax and abdomen blocks.

Graphs and statistical analysis
The internal consistency of the post-intervention survey was assessed using Cronbach's alpha.There were two question sets that evaluated different themes.These included anatomical and imaging knowledge in one set as well as confidence using Pacsbin and identification of anatomical structures of the thorax and abdomen in another set.Cronbach's alpha was determined using Microsoft Excel.
The formula used to solve for Cronbach alpha is provided below: K equals the number of questions in the survey.Sy 2 is the variance of the total score.The Sum Si 2 is sum of the variance for each question item.Values corresponding to 0.7 to 1 are considered acceptable.
GPower version 3.1.9.7 was used for the statistical power calculations and sample size needed to achieve statistical significance.To calculate statistical power achieved, a post hoc calculation using a two-sided T test for differences of means was used.An alpha rate of 0.05 and the sample sizes for the student groups were given to the program.From here, the effect sizes were calculated and were used to determine statistical power.Similarly, to compute the sample size needed to achieve statistical significance, a two-sided T test a priori calculation was performed in which the alpha rate of 0.05, the power of 0.8 and the effect size were used.

RE SULTS Preintervention findings from student participants
We conducted a preintervention questionnaire to obtain information regarding participants' prior anatomical and imaging experience in addition to demographic information and major in college.Out of the 57 participants, 47 responded (82% response rate) to the preintervention questionnaire.

Participants' anatomy and imaging experience
Participants were asked whether they had anatomy experience prior to matriculation at LCOM.A total of 22 participants (47%) indicated prior anatomical experience and the remaining 25 (53%) indicated no prior anatomical experience.Of the participants who indicated prior anatomical experience, about half reported some form of anatomy undergraduate or graduate lecture course while the other half reported undergraduate/graduate lecture with a dissection experience involving human cadavers.Participants were also queried about their imaging experience prior to matriculation at LCOM.Fewer participants, at a total of 15 (32%), reported prior imaging experience while the remaining 32 (68%) participants indicated no prior imaging experience.Of those with prior imaging experience, X-ray, CT, and MRI were the most frequently cited modalities.Next, participants were given the opportunity to self-assess their anatomical and imaging knowledge prior to their matriculation at LCOM.A total of 45 and 46/57 (78% and 81%) responses were obtained for both questions.Most participants selected their anatomical knowledge to be good or very good (71%) and their imaging knowledge to be fair or good (72%).The results showed that more participants reported pre-matriculation anatomical experience as opposed to imaging experience and as a result, likely rated their anatomical knowledge better than their imaging knowledge.
Participants' college major/minor Participants were asked to disclose their major/minor while in college.
Most participants at 31 individuals disclosed a major in the biological sciences, while 5 individuals reported majors in engineering, chemistry, and mathematics, and 4 individuals reported a major in the humanities.

CT scans from body donor sources
Since this study incorporated scans from post-mortem body donors that had been previously embalmed, the authors sought to evaluate the presence of artifacts of fixation present in the images.To illustrate this, a representative axial scan of the thorax is shown from a body donor in which fluid can be seen in the pleural cavities as well as coagulated blood in the left atrium (Figure 2).While artifacts of the tissue changes were present, there was evidence of medical devices that were visible on some scans.In the same image, an aortic valve stent, which appeared as a hyper-dense ring around the aortic root, can be seen in addition to cardiac pacemaker leads and coronary stents (Figure 2).Other body donors exhibited fluid in body cavities as well as coagulated blood in heart chambers and had similar medical devices which showed up on scans.

Intergroup performance
Differences in group performance on practical anatomy examinations were evaluated to determine whether the source of scans had an impact on performance.Anatomical spatial knowledge was assessed using questions which asked the participant about an anatomical relationship (e.g.what muscle is found deep to the tagged structure, which vessel runs with this nerve, etc.).Differences in academic performance between groups and within groups were evaluated for the thorax and abdomen/pelvis blocks of anatomy.Overall, there were no statistically significant differences in practical examination scores, ability to correctly identify anatomical structures on CT images, and spatial relationship knowledge between the three study groups (Figure 3).Interestingly, all participants performed well on CT-based image questions, scoring a 70% or higher in both the thorax and abdomen/pelvis blocks of anatomy (Figure 3).For bonus questions, one CT-based question was included on the abdomen/ pelvis practical examination which tested retention of material from the thorax section and another bonus question was included in the head/neck practical which tested retention of material from the abdomen/pelvis block.For both bonus questions, there were no statistically significant differences between the study groups (Figure 3).A Cohen's d was performed for the thorax and abdomen/pelvis practical examination scores and results demonstrated values lower than 0.3, illustrating a small effect size between the participants in the three groups. 16

Intragroup performance
With respect to intragroup differences (e.g.Group 1 performance on the thorax versus abdomen/pelvis practical examinations), there were no statistically significant differences in practical examination scores as well as performance on spatial relationships and CT image questions (Figure 4).Interestingly, participants in all study groups performed significantly better on abdomen/pelvis spatial anatomical relationship questions as opposed to thorax spatial anatomical relationship questions (Figure 4).The achieved power for differences in the thorax practical was 0.12-0.15.For the abdomen the achieved power was 0.05-0.07.Both values are much lower than the 0.8 (e.g.80%) power needed to detect significant differences.
Together, these results demonstrate that, while there were no significant differences in academic performance between the study groups, significant differences were observed with respect to anatomical spatial knowledge within each study group.

Anatomy practical examination performance between 2021 and 2022
Differences in anatomy practical examination performance on content in the thorax and abdomen/pelvis blocks was evaluated between students taking FoCS in 2021 and 2022.The objective was to determine whether the imaging component in the anatomy laboratory had any noticeable effect on students' practical examination performance.After plotting the mean anatomy practical examination data for the thorax and abdomen/pelvis blocks, Post-mortem findings on a body donor scan.Relevant anatomical structures on this axial section have been shown.Note the presence of fluid in the pleural cavities of the body donor, a result of the embalming process.Significant clotting of blood was also evident in the large vessels as well as the chambers of the heart in this body donor scan.An aortic valve stent as well as coronary stents and cardiac pacemaker leads can be seen on the scan.
there were no significant differences between the 2021 and 2022 academic years (Figure 5), suggesting that the intervention neither enhanced nor hindered anatomy practical examination performance for these two blocks of anatomy.

Participant perceptions of the intervention
At the end of the intervention, participants were asked to respond to an 18-item questionnaire designed to assess participant perceptions about the CT intervention.

Participants' confidence using Pacsbin
In the first set of survey questions, participants' confidence using Pacsbin and ability to correctly identify anatomical structures from a stack of patient images was queried.It was anticipated that all participants, regardless of group assignment, would agree with these items following the intervention.As expected, most participants selected somewhat confident on a 4-point Likert scale in response to these question (Figure 6A).When asked if they believed the CT intervention improved their knowledge of anatomy and ability to identify anatomical structures on cross-sectional images, most participants selected strongly agree/agree in response to both items (Figure 6B).

Participants' perceptions on preparedness for imaging in clerkships
Next, participants were asked if they believed the intervention would better prepare them for use of imaging software andinterpretation of imaging findings upon entering clerkships.Here, most participants agreed with these questions (Figure 7A).

Participants' perceptions on pathology knowledge and donor humanity
Since participants from Group 1 were more likely to visualize pathology/medical devices in their body donor during dissection as well as on their corresponding scans, the authors were curious whether the intervention improved Group 1 participants' knowledge of pathology and supported a better appreciation for their body donor's humanity.With regard to pathology knowledge, responses F I G U R E 3 Intergroup differences on anatomy practical examinations.Participants in each group were assessed for differences in examination performance in the Thorax and Abdomen/Pelvis blocks of the course.There were no significant differences between the groups in any category evaluated.Graphs labeled Anatomy Practicals include results from questions on body donors, prosections, bones and CT images.Graphs labeled CT ID include results from identification (part A) questions on CT scans.Graphs labeled CT include results on both parts A (identification) and B (application) questions associated with the two CT images.Graphs labeled Spatial Anatomical Knowledge include performance on questions that assessed spatial relationships.Graphs labeled Bonus CT include questions which evaluated performance on both parts A (identification) and B (application) questions associated with material from the previous block.Data are presented as the mean ± the SD.
were largely split between agree and neither agree nor disagree (Figure 7B).There were no group differences in responses.When asked whether the CT intervention supported a better appreciation for the body donors' humanity, most participants indicated neither agree nor disagree with participants in Group 2 agreeing at a higher rate compared to participants from Groups 1 and 3 (Figure 7C).

Participants' perceptions on dissections
Participants were asked whether the intervention helped them conduct their dissections more confidently and efficiently.It was anticipated that participants in Group 1 would be more likely to agree with these items considering that they could view scans from their body donor in preparation for assigned dissections as well as during dissections.Surprisingly, most participants selected neither agree nor disagree and disagree to these questions (Figure 8A,B).
Looking at group differences in responses, it was observed that participants in Group 2 tended to agree at a higher rate on these two items (Figure 8A,B).

Participants' perceptions on anatomical and imaging knowledge
Since participants' perceived anatomical and imaging knowledge was assessed before and after the intervention, the results from these questions were compared to determine if participants rated their anatomical and imaging knowledge higher post-intervention.
As expected, participants rated their anatomical and imaging knowledge better with more responses in the excellent and very good categories for anatomy and the very good and good categories for imaging (Figure 9).Participants were asked to disclose what resources they commonly used (if any) to supplement their CT F I G U R E 4 Intragroup differences on anatomy practical examinations.Participants in each group were evaluated on differences in their performance between Thorax and Abdomen/Pelvis practicals.There were no differences in any category evaluated with the exception of spatial anatomical knowledge where significant differences were observed for each group between the thorax and abdomen/pelvis practical examinations.Graphs labeled Anatomy Practicals include results from questions on body donors, prosections, bones and CT images.Graphs labeled CT ID include results from identification (part A) questions on CT scans.Graphs labeled CT include results on both parts A (identification) and B (application) questions associated with the two CT images.Graphs labeled Spatial Anatomical Knowledge include performance on questions that assessed spatial relationships.Graphs labeled Bonus CT include questions which evaluated performance on both parts A (identification) and B (application) questions associated with material from the previous block.Data are presented as the mean ± the SD.** ≤ 0.01; **** ≤ 0.0001.
learning as well as how often they utilized their assigned images.
As shown in Table 1, more participants in Group 1 used outside resources as compared to participants in Groups 2 and 3. Webbased resources and applications were commonly listed as resources used to supplement CT learning.Most participants indicated that they used their assigned images in the Thorax and Abdomen blocks a few times during the course of the study (Table 2).
Cronbach's alpha was used to assess the internal consistency of questions in the post-intervention survey.Here, questions 5-8 and 14-15 (question set #1) as well as questions 9-11 (question set #2) were evaluated separately as these items assessed different themes including anatomical and imaging knowledge and confidence using Pacsbin and with anatomical identification on patient images, respectively.Results showed Cronbach's alpha coefficients of 0.82 and 0.74 indicating an acceptable reference range (0.7-1) and thus reliability of the survey tool.
Taken together, these findings support the positive impact that the intervention had on participants' perceptions about their anatomical and imaging knowledge and highlight some group differences with regard to perspectives.

DISCUSS ION
This study illustrates two major findings.First, the addition of CT imaging to the anatomy curriculum supported the authors' curricular objectives which were to: 1. Increase early exposure to crosssectional imaging, a goal that was met through the implementation of the imaging intervention in the thorax and abdomen/pelvis blocks of anatomy; 2. improve anatomical knowledge on cross-sectional images, a goal that was achieved based on participant performance on CT image based questions from practical examinations as well as Anatomy practical examination scores between years.Student performance on anatomy practical examinations in the Thorax and Abdomen/Pelvis blocks was evaluated between 2021 and 2022 academic years.There were no significant differences in practical examination performance between these years.Data are presented as the mean ± the SD.

F I G U R E 6
Post-intervention survey-confidence and knowledge assessments.Participants were queried on a 4-point or 5-point Likert scale to assess perceptions on confidence using Pacsbin and general anatomical knowledge.(A) Most participants indicated that they felt somewhat confident using Pacsbin and identifying anatomical structures in the thorax and abdomen if presented with a stack of patient images.(B) Most participants selected strongly agree or agree in response to the question asking whether learning the cross-sectional anatomy aided anatomical knowledge and ability to identify anatomical structures on scans.
participant survey results; and 3. improve student confidence using imaging software, achieved based on participant survey results.
Second, there were no major differences in academic performance between groups using scans from body donors or from living individuals, supporting the hypothesis that significant differences between group scores would not be apparent.
One of the novel features of the study was the mode by which participants were introduced to CT images and then tested on CT F I G U R E 7 Post-intervention survey-beliefs assessments.Participants were queried using yes/no/neutral questions about their preparedness for use of imaging in clerkships as well as their beliefs regarding pathology and humanity.(A) Most participants selected yes in response to questions asking about preparedness for the use of imaging software and interpretation of imaging findings upon entering clerkships.(B) Participants responses were primarily split between agree and disagree with the statement asking whether the intervention improved understanding of pathology.There were no differences in group perceptions on this question.(C) Participants were primarily neutral when asked whether the intervention supported a better appreciation for their cadaver's humanity.There was a significant difference between Groups 1 and 2, with participants in Group 2 tending to agree more with the statement that the intervention supported a better appreciation for the cadaver's humanity.Data are presented as the mean ± the SD.Statistical significance was considered at *p ≤ 0.05.An important component of this study was to investigate whether use of scans from body donors or from living subjects best supported the curricular needs which were to enhance students' anatomical and imaging knowledge as well as confidence using Pacsbin.The results show there were no differences in academic performance between participants studying scans from body donors versus scans from living subjects.0]12,19 For example, Lufler et al. 9 found that firstyear medical students who studied cadaveric CT scans scored higher on anatomy practical examinations, performed better on questions associated with spatial anatomy and received higher overall course grades than students who did not use CT scans.
Since Lufler et al. 9 made review of cadaver CT images optional, it is possible that the students who opted to study from CT scans were more intrinsically motivated to do additional studying, and therefore more likely to perform better on examinations.Indeed, it is possible that the students who elected to participate in the current study were more intrinsically motivated, a factor which Comparison of anatomy and imaging knowledge pre versus post-intervention.Participants were queried on a 5-point Likert scale about their perceptions surrounding their anatomical and imaging knowledge prior to and after the intervention.There were more participants who selected their anatomical knowledge to be excellent or very good and more participants who selected their imaging knowledge to be very good or good in the post-intervention questionnaire.
TA B L E 1 Use of outside imaging resources.may explain why we did not observe any statistically significant differences in academic performance between the three groups.A more recent study also concluded that incorporating cadaveric CT scans into the general anatomy curriculum significantly improved student test scores in anatomy. 12In contrast, Murakami et al. 11 found that the introduction of post-mortem CT scans did not improve medical students' academic performance in radiologybased questions or anatomy.The findings in this study align with those of Murakami et al. 11 and additionally support the hypothesis that major differences in academic performance would not be apparent between the three groups.Although the hypothesis was upheld, it is still important to explore the reasons associated with the lack of differences.For example, it is possible that these observations were in part due to participants use of outside resources.Indeed, participants in all groups cited that they used outside resources to supplement their learning.This may have been particularly advantageous to those individuals who had been assigned cadaveric scans.[22] Specifically, O'Donnell and Woodford 21 cited tissue decomposition and pathological findings in addition to lack of contrast agent in post-mortem scans as a challenge for students trying to differentiate normal post-mortem findings from a premortem illness.
Similarly, Slon et al. 22 discovered that post-embalming artifacts led to suboptimal learning experiences for students.The authors circumvented this challenge by obtaining images of cadavers from local hospitals and clinics taken prior to their death. 22As cadavers have been and continue to be used in imaging education, it is important to consider whether ante-mortem images can be obtained or what conditions best support tissue preservation for imaging purposes.Considering that the first-year medical student is likely a novice learner, the authors advise that use of images from healthy, living individuals would best support students' ability to identify and learn the anatomy.With time, cadaveric images could be introduced into the curriculum to further support students' ability to integrate anatomical and pathology knowledge with imaging findings.
In addition to the impact that cadaveric imaging has on student anatomical knowledge, several groups have reported improvements in students' understanding of spatial anatomical relationships after studying body donor CT images. 9,12,22,23In this manner, the authors of the current study expected that spatial anatomical knowledge would be superior for participants reviewing scans from their body donor (Group 1) as opposed to participants reviewing scans from other body donors (Group 2) or living subjects (Group 3).The rationale was that participants using body donor scans would be better able to connect the 3D anatomy during dissection with the 2D anatomy in the scans and would therefore have a better knowledge of spatial relationships.The authors did not observe this correlation in this study.Participants in all three groups performed similarly on questions related to spatial relationships.It is possible that the source (e.g.cadaveric versus living) of imaging is not as important as the inclusion of imaging in general to guide student understanding of spatial relationships.
Indeed, others have reported positive benefits associated with use of nonpathological CT images from living subjects to enhance students' knowledge of spatial relationships, 24,25 further supporting the importance of imaging in general for spatial anatomical understanding.A recent study took CT visualization one step further by asking whether the modality of viewing the images made a difference with respect to spatial understanding. 23Here, medical students were placed into two groups: those who viewed cadaveric CT images using augmented reality, providing a more immersive 3D environment, and those who viewed cadaveric images on a traditional 2D computer screen. 23Those students who used the augmented reality performed better on post-test questions which investigated understanding of spatial relationships, a finding the authors attribute to the stereoscopic depth cues afforded by augmented reality. 23As technology continues to become incorporated into medical education curriculum, it may be important to consider use of immersive environments for supporting students' knowledge of imaging and anatomical spatial relationships.
Despite the lack of differences in academic performance, the authors did find that participants' overall satisfaction and beliefs . 12One surprising finding from the questionnaire data was that participants overall did not report a greater appreciation for their body donor's humanity.This is in contrast to a report by McBain et al. 10 who documented that medical students who reviewed CT images from body donors being dissected as opposed to students reviewing pathological images of a body donor not being dissected exhibited a more humanistic approach to the dissection experience, identifying the body donor as human as opposed to a specimen.One key difference is that the students in the McBain et al. 10 study were fourth-year medical students who were learning clinical vignettes from premortem imaging of their cadavers that contained pathological findings.It is likely that given the stage of training that the students were in, they had prior exposure to pathology in the clinical setting, and as a result, may have been better able to identify with their body donor as a patient with a medical history.In the current study, it is possible introduced premortem CT scans of cadavers to fourth-year medical students participating in an elective course: Anatomy for Surgeons at McGill University.
Review Board, Ethics Committee on Human Research at the University of Vermont approved this study: IRB approval (STUDY00002139).All first-year medical students enrolled in the Foundations of Clinical Sciences (FoCS) in 2022 were informed of the study, its goals and use of participant data.Participation was voluntary and students could opt out at any time.Students were requested to review the consent form and sign it prior to participation in the study.For use of de-identified cadaveric CT images, permission was obtained from the Anatomical Gift Program at the UVM LCOM.

FoCS, which introduces
students to many disciplines in the basic sciences including anatomy, physiology, imaging, metabolism, histology, genetics, and ethics.The anatomy component of the course is introduced in 6 blocks.Each block focuses on regional anatomy and includes: Back (Block 1), Upper Limb (Block 2), Lower Limb (Block 3), Thorax (Block 4), Abdomen and Pelvis (Block 5), and Head and Neck (Block 6).Anatomical instruction involved laboratory dissections, electronic modules (e.g.recorded lectures) that students completed at their own pace, and faculty-guided clinical correlations sessions, intended to help students apply anatomical knowledge to clinical reasoning.For anatomical dissection, one-third of the class alternated dissections.Here, 124 students were separated into 3 groups: Groups A, B, and C. Each group consisted of 41-42 students.Each table had 4-5 students who participated in any one dissection.Three faculty and 1-2 teaching assistants (TAs) were available in each laboratory session to assist students with their dissections.Each laboratory session was 3 h long and students were encouraged to prepare for their dissection by reviewing their dissector instructions (Bernd Dissector, a virtual photographic dissector developed by Paulette Bernd, Anatomy Professor, Columbia University), watch the accompanying dissection video (made in house by the anatomy faculty at LCOM in 2020), and review the dissection checklist, which contained the list of structures to be found along with relevant anatomical information.Upon arriving in the laboratory, students were encouraged to review dissections from previous groups prior to engaging in their assigned dissection.Students had access to the prosected cadaver as well as additional prosected specimens for review.There were a total of 29 laboratory sessions broken down as follows: three laboratory sessions in Block 1 (Back); six in Block 2 (Upper Limb), five in Block 3 (Lower Limb), four in Block 4 (Thorax), five in Block 5 (Abdomen and Pelvis), including a faculty-guided tour of the pelvic anatomy, and six in Block 6 (Head and Neck) including a faculty-guided tour of the cranial nerves and calvaria.With the exceptions of the two faculty-guided tours in Blocks 5 and 6, students participated in a total of 9 dissections.Assessments in anatomy included practical laboratory examinations which occurred for each block of anatomy.Here, structures were tagged on cadavers, prosections, bones and where applicable, CT images.Students were asked to (A) identify the structure and (B) describe something about the structure (i.e.what is the function of this structure, name one vascular branch of this structure, etc.).

4 .
Better prepare students for interpreting diagnostic imaging in the clinical environment.
All information from questionnaires was obtained using Qualtrics XM, version 2020, (Qualtrics, Provo, UT).Participant responses were saved in a Microsoft Excel spreadsheet (Microsoft, Redmond, WA).Data in the Excel spreadsheet was imported into GraphPad Prism, version 9.0 for Windows (GraphPad Prism Software, San Diego, CA) which was used to generate graphs and perform statistical analyses.Results were presented as the percent and sum of responses for questionnaire data or the mean ± the standard deviation (±SD) for examination data.For analyses of intra-group performance (e.g.Group 1 practical exam scores from the Thorax and Abdomen) and performance between academic years 2021 and 2022, a students' unpaired t-test was used to evaluate statistical significance.For analyses of intergroup performance (e.g.Groups 1, 2 and 3 on the Thorax practical), a one-way ANOVA with Turkey's post-test was used to evaluate statistical significance.A p value of ≤0.05 was considered significant.To evaluate the effect size, a Cohen's d was performed using Microsoft Excel.Here, the means and standard deviations of the practical exam scores for the thorax and abdomen were calculated for each group.Cohen's d was analyzed based on the mean of one group subtracted from the mean of another group divided by their pooled standard deviations.The formula for determining the pooled standard deviation is shown below: N 1 is the number of participants in Group 1, and N 2 is the number of participants in Group 2. SD 1 is the standard deviation of values for participants in Group 1, and SD 2 is the standard deviation of values for participants in Group 2.

F I G U R E 8
Post-intervention survey-dissection assessments.Participants were queried on a 5-point Likert scale to assess to what extent they agreed with the statement that the intervention supported more confidence and efficiency during dissections.(A) Most participants were neutral or disagreed with the statement that the intervention supported more confidence during dissections.Participants in Group 2 tended to agree at a significantly higher rate to this statement than with participants in Groups 1 and 3. (B) Most participants were neutral or disagreed with the statement that the intervention supported more efficiency during dissection.While participants in Group 2 tended to agree more with this statement, the results are not statistically significant.Data are presented as the mean ± the SD.Statistical significance was considered at *p ≤ 0.05.image questions during anatomy practical examinations.In this study, Pacsbin was used as a DICOM viewer.Here, participants not only visualized their scans on iPads while in the laboratory but were tested on structures using Pacsbin as well.For examinations, participants were allowed to view the labeled structures in the axial, coronal, and sagittal planes and could additionally scroll through the image slices to view the progression of the structures.This modality of testing is more representative of how students will visualize CT images upon entering the clinical environment and contrasts with the more traditional use of static images for testing.17,18Another novel component of this project was the interdisciplinary nature by which the study was constructed as well as the learning materials provided to the students.With respect to learning materials, short instructional modules were created by a diagnostic radiologist, while radiology residents provided laboratory-based assistance with structure identification on CT images.Senior medical students provided supplemental classroom-based learning sessions to guide student engagement with Pacsbin and labeling of anatomical structures.While the authors envisioned that participants would label anatomical structures while in the anatomy laboratory, the annotation feature on Pacsbin was not compatible with the iPad.Despite this drawback, several students engaged in these supplemental classroom sessions and sought feedback from senior medical students and faculty regarding the accuracy of their labeling.Senior medical students additionally constructed quizzes in Pacsbin that were specific to each study group.Together, these approaches helped participants engage with the learning materials, likely contributing to high participant satisfaction with the intervention.

TA B L E 2
Participants' responses by group for the use of outside resources to supplement learning of cross-sectional anatomy on CT scans.Frequency of use of CT images.Note: Participants' responses by group for the frequency of use of assigned CT images.
about their anatomical knowledge and confidence with the use of Pacsbin were positive following the intervention.Specifically, participants reported a perceived benefit of this intervention toward future use of imaging software and interpretation of imaging findings in the clinical environment and additionally felt that their anatomical and imaging knowledge were enhanced following this intervention.Others have similarly documented favorable participant perceptions following anatomical imaging interventions.For example, Murakami et al.11 reported that medical students generally regarded cadaveric CT images positively as it pertained to early exposure to CT imaging.Others found that the introduction of imaging early in medical student training enhanced medical students' confidence in the use of standard DICOM viewers and improved students' perceptions on their abilities to identify major anatomical structures in stacks of CT images(Wilson et al.,   2017)