<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD Journal Publishing DTD v2.3 20070202//EN" "journalpublishing.dtd">
<article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" article-type="brief-report" dtd-version="2.3" xml:lang="EN">
<front>
<journal-meta>
<journal-id journal-id-type="publisher-id">Front. Vet. Sci.</journal-id>
<journal-title>Frontiers in Veterinary Science</journal-title>
<abbrev-journal-title abbrev-type="pubmed">Front. Vet. Sci.</abbrev-journal-title>
<issn pub-type="epub">2297-1769</issn>
<publisher>
<publisher-name>Frontiers Media S.A.</publisher-name>
</publisher>
</journal-meta>
<article-meta>
<article-id pub-id-type="doi">10.3389/fvets.2024.1437284</article-id>
<article-categories>
<subj-group subj-group-type="heading">
<subject>Veterinary Science</subject>
<subj-group>
<subject>Perspective</subject>
</subj-group>
</subj-group>
</article-categories>
<title-group>
<article-title>Role of AI in diagnostic imaging error reduction</article-title>
</title-group>
<contrib-group>
<contrib contrib-type="author" corresp="yes">
<name><surname>Burti</surname> <given-names>Silvia</given-names></name>
<xref ref-type="corresp" rid="c001"><sup>&#x002A;</sup></xref>
<uri xlink:href="https://loop.frontiersin.org/people/1025630/overview"/>
<role content-type="https://credit.niso.org/contributor-roles/writing-review-editing/"/>
</contrib>
<contrib contrib-type="author">
<name><surname>Zotti</surname> <given-names>Alessandro</given-names></name>
<uri xlink:href="https://loop.frontiersin.org/people/2801312/overview"/>
<role content-type="https://credit.niso.org/contributor-roles/writing-review-editing/"/>
</contrib>
<contrib contrib-type="author">
<name><surname>Banzato</surname> <given-names>Tommaso</given-names></name>
<uri xlink:href="https://loop.frontiersin.org/people/767486/overview"/>
<role content-type="https://credit.niso.org/contributor-roles/writing-original-draft/"/>
<role content-type="https://credit.niso.org/contributor-roles/writing-review-editing/"/>
</contrib>
</contrib-group>
<aff><institution>Department of Animal Medicine, Production and Health, University of Padua</institution>, <addr-line>Padua</addr-line>, <country>Italy</country></aff>
<author-notes>
<fn id="fn0001" fn-type="edited-by"><p>Edited by: Hussein M. El-Husseiny, Tokyo University of Agriculture and Technology, Japan</p></fn>
<fn id="fn0002" fn-type="edited-by"><p>Reviewed by: Elissa Randall, Colorado State University, United States</p>
<p>Sofia Alves-Pimenta, Universidade de Tr&#x00E1;s-os-Montes e Alto, Portugal</p></fn>
<corresp id="c001">&#x002A;Correspondence: Silvia Burti, <email>silvia.burti@unipd.it</email></corresp>
</author-notes>
<pub-date pub-type="epub">
<day>30</day>
<month>08</month>
<year>2024</year>
</pub-date>
<pub-date pub-type="collection">
<year>2024</year>
</pub-date>
<volume>11</volume>
<elocation-id>1437284</elocation-id>
<history>
<date date-type="received">
<day>23</day>
<month>05</month>
<year>2024</year>
</date>
<date date-type="accepted">
<day>21</day>
<month>08</month>
<year>2024</year>
</date>
</history>
<permissions>
<copyright-statement>Copyright &#x00A9; 2024 Burti, Zotti and Banzato.</copyright-statement>
<copyright-year>2024</copyright-year>
<copyright-holder>Burti, Zotti and Banzato</copyright-holder>
<license xlink:href="http://creativecommons.org/licenses/by/4.0/">
<p>This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.</p>
</license>
</permissions>
<abstract>
<p>The topic of diagnostic imaging error and the tools and strategies for error mitigation are poorly investigated in veterinary medicine. The increasing popularity of diagnostic imaging and the high demand for teleradiology make mitigating diagnostic imaging errors paramount in high-quality services. The different sources of error have been thoroughly investigated in human medicine, and the use of AI-based products is advocated as one of the most promising strategies for error mitigation. At present, AI is still an emerging technology in veterinary medicine and, as such, is raising increasing interest among in board-certified radiologists and general practitioners alike. In this perspective article, the role of AI in mitigating different types of errors, as classified in the human literature, is presented and discussed. Furthermore, some of the weaknesses specific to the veterinary world, such as the absence of a regulatory agency for admitting medical devices to the market, are also discussed.</p>
</abstract>
<kwd-group>
<kwd>artificial intelligence</kwd>
<kwd>error</kwd>
<kwd>machine learning</kwd>
<kwd>image quality</kwd>
<kwd>radiology&#x2014;education</kwd>
</kwd-group>
<counts>
<fig-count count="0"/>
<table-count count="1"/>
<equation-count count="0"/>
<ref-count count="31"/>
<page-count count="4"/>
<word-count count="3614"/>
</counts>
<custom-meta-wrap>
<custom-meta>
<meta-name>section-at-acceptance</meta-name>
<meta-value>Veterinary Imaging</meta-value>
</custom-meta>
</custom-meta-wrap>
</article-meta>
</front>
<body>
<sec sec-type="intro" id="sec1">
<label>1</label>
<title>Introduction</title>
<p>The topic of error mitigation in diagnostic imaging is a relatively unexplored field in the veterinary literature. Indeed, to the best of the authors&#x2019; knowledge, only two papers investigating such a topic are available (<xref ref-type="bibr" rid="ref1">1</xref>, <xref ref-type="bibr" rid="ref2">2</xref>). Likewise, incidence rates and the overall costs associated with diagnostic imaging errors have been poorly investigated in veterinary medical practice. Indeed, only one study (<xref ref-type="bibr" rid="ref3">3</xref>) reports the radiologic error rate being comparable to what is reported in human medicine. Instead, an entire set of literature devoted to analyzing the most common causes of diagnostic imaging errors, along with possible solutions, is currently available in human medicine (<xref ref-type="bibr" rid="ref4">4</xref>, <xref ref-type="bibr" rid="ref5">5</xref>). It is important to understand that diagnostic imaging errors are much more intricate than they might seem because they involve a complex interaction between individual psychological (<xref ref-type="bibr" rid="ref6">6</xref>) environmental, and educational factors (<xref ref-type="bibr" rid="ref7">7</xref>). A diagnostic error is defined as a &#x201C;deviation from the expected norm&#x201D; (<xref ref-type="bibr" rid="ref8">8</xref>), and the consequences for the patient may vary from no consequences to death. Renfrew et al. (<xref ref-type="bibr" rid="ref9">9</xref>) first proposed a comprehensive classification of the causes of diagnostic imaging errors, which were subsequently modified by Kim and Mansfield (<xref ref-type="bibr" rid="ref10">10</xref>). In addition, some authors have approached this complex theme from different perspectives, ranging from the identification of different cognitive biases (<xref ref-type="bibr" rid="ref6">6</xref>), to the analysis of interpretative errors (<xref ref-type="bibr" rid="ref4">4</xref>), to the strategies for error reduction (<xref ref-type="bibr" rid="ref11">11</xref>).</p>
<p>It is important to note at this point that a universally recognized &#x201C;etiology&#x201D; of errors in human diagnostic imaging is currently unavailable, and the definitions and the solutions proposed for different scenarios may vary among authors. In recent years, we have witnessed an increased interest in the applications of AI in the veterinary diagnostic imaging field (<xref ref-type="bibr" rid="ref12">12</xref>, <xref ref-type="bibr" rid="ref13">13</xref>). Among other applications, AI is mainly used as a supportive tool to guide the interpretation of medical images in veterinary medicine. Even if AI is reported to have an overall lower error rate than radiologists have both in human (<xref ref-type="bibr" rid="ref14">14</xref>) and veterinary medicine (<xref ref-type="bibr" rid="ref15">15</xref>), dealing with such a technology is not as straightforward as it might seem (<xref ref-type="bibr" rid="ref16">16</xref>, <xref ref-type="bibr" rid="ref17">17</xref>). This perspective analysis aims to examine the role of AI in mitigating each source of error in veterinary imaging through the error classification suggested by Kim and Mansfield (<xref ref-type="bibr" rid="ref10">10</xref>).</p>
</sec>
<sec id="sec2">
<label>2</label>
<title>Types of errors and role of AI in mitigation</title>
<sec id="sec3">
<label>2.1</label>
<title>Complacency</title>
<p>&#x201C;Complacency refers to over-reading and misinterpretation of findings, a finding is detected but attributed to the wrong cause (false positive-error)&#x201D; (<xref ref-type="bibr" rid="ref10">10</xref>). This type of error is reported to be uncommon (0.9%) in human medicine, whilst no data are available in veterinary medicine. In this latter field, a discrepancy between the AI system output and the radiologist&#x2019;s interpretation is likely to occur. AI systems are reported to generate lower error rates (including both false positives and false negatives) than radiologists (at least for some specific findings) (<xref ref-type="bibr" rid="ref15">15</xref>). Veterinary radiologists should therefore consider reinterpreting findings, taking the AI results into account.</p>
</sec>
<sec id="sec4">
<label>2.2</label>
<title>Faulty reasoning</title>
<p>&#x201C;Error of over-reading and misinterpretation, in which a finding is appreciated and interpreted as abnormal but is attributed to the wrong cause. Misleading information and a limited differential diagnosis are included in this category&#x201D; (<xref ref-type="bibr" rid="ref10">10</xref>). At present, the available AI systems only detect specific radiographic findings (<xref ref-type="bibr" rid="ref15">15</xref>, <xref ref-type="bibr" rid="ref18">18</xref>, <xref ref-type="bibr" rid="ref19">19</xref>) and are not able to provide differential diagnoses based on the clinical findings. Large language models (LLMs) (<xref ref-type="bibr" rid="ref20">20</xref>) capable of interpreting the images and generating a list of differentials based on the medical history will soon be available, thus potentially reducing this type of error.</p>
</sec>
<sec id="sec5">
<label>2.3</label>
<title>Lack of knowledge</title>
<p>&#x201C;The finding is seen but is attributed to the wrong cause because of a lack of knowledge on the part of the viewer or interpreter&#x201D; (<xref ref-type="bibr" rid="ref10">10</xref>). This type of error is, to the authors&#x2019; knowledge, particularly relevant in the veterinary scenario, where most radiographic images are not interpreted by a radiologist but by general practitioners. As mentioned earlier, current AI-based systems cannot correlate the imaging findings with a specific list of differentials based on the medical history and therefore, to date, AI has had limited impact in mitigating this type of error.</p>
</sec>
<sec id="sec6">
<label>2.4</label>
<title>Under-reading</title>
<p>&#x201C;The lesion is not detected.&#x201D; According to Kim and Mansfield (<xref ref-type="bibr" rid="ref10">10</xref>), this alone accounts for 42% of the total diagnostic errors. Under-reading is, most likely, one of the main reasons for implementing AI systems in the day-to-day routine. Indeed, under-reading stands as a very common problem that might arise from both individual and environmental situations (<xref ref-type="bibr" rid="ref7">7</xref>). The role of AI in mitigating this type of error is, potentially, a game changer as AI systems are not subjected to cognitive biases or environmental contexts (overworking, challenging working environment, distractions, etc.). On the other hand, the final user needs to consider that AI system accuracy is also affected by several factors, such as image quality or lesion rate in the database (<xref ref-type="bibr" rid="ref21">21</xref>). Lastly, the user needs to be aware that most of the veterinary AI-based systems have a variable reported accuracy in the detection of specific lesions. For instance, accuracy in detecting pleural effusion is usually very high (<xref ref-type="bibr" rid="ref15">15</xref>, <xref ref-type="bibr" rid="ref18">18</xref>, <xref ref-type="bibr" rid="ref22">22</xref>) whereas accuracy for pulmonary nodules or masses is significantly lower (<xref ref-type="bibr" rid="ref18">18</xref>, <xref ref-type="bibr" rid="ref23">23</xref>).</p>
</sec>
<sec id="sec7">
<label>2.5</label>
<title>Poor communication</title>
<p>&#x201C;The lesion is identified and interpreted correctly, but the message fails to reach the clinician.&#x201D; Reliable communication of imaging findings is vital for the correct management of patients, both in veterinary and human medicine. Imaging reports use highly specialized terminology, and the accurate interpretation of these terms relies on the expertise of the referring clinician. This type of error is reported to be quite rare (<xref ref-type="bibr" rid="ref10">10</xref>) as, when a report is unclear, a direct explanation is usually required from the reporting physician. To this end, incorporating large language models (LLMs) (<xref ref-type="bibr" rid="ref20">20</xref>) within the reporting systems could help in creating more homogeneous reports and therefore improve communication between the clinician and the radiologist.</p>
</sec>
<sec id="sec8">
<label>2.6</label>
<title>Prior examination/history</title>
<p>&#x201C;The finding is missed because of failure to consult prior radiologic studies or reports&#x201D; and &#x201C;The finding is missed because of the acquisition of inaccurate or incomplete clinical history.&#x201D; These are among the most common types of errors, and the American College of Radiology recommends that all the patients&#x2019; previous reports should be available to the radiologist during exam evaluation (<xref ref-type="bibr" rid="ref10">10</xref>). This type of error is most relevant in teleradiology services since most of these services do not have access to complete patient history. AI-based products guiding radiologists (both in human and veterinary medicine) throughout the reporting process (from image acquisition to final report) could be important in mitigating these errors. For example, using LLMs to promptly summarize the patient&#x2019;s clinical history could provide the radiologist with quick and useful information.</p>
</sec>
<sec id="sec9">
<label>2.7</label>
<title>Location</title>
<p>&#x201C;The finding is missed because the location of a lesion is outside the area of interest on an image, such as in the corner of an image.&#x201D; These errors are fairly common and are possibly related to what is referred to as &#x201C;intentional&#x201D; or &#x201C;tunnel vision bias&#x201D; (<xref ref-type="bibr" rid="ref10">10</xref>). These are well-known cognitive biases. In a famous experiment, radiologists were asked to detect pulmonary nodules from CT images. The picture of a gorilla, 10 times larger than the average nodule, was placed in one of the CT images. Surprisingly, 83% of the radiologists did not report seeing the gorilla, despite eye-tracking technologies demonstrating that all the radiologists looked at it (<xref ref-type="bibr" rid="ref24">24</xref>). In this case, using AI systems to assist the radiologist could help in reducing these types of errors provided that the AI systems themselves do not generate numerous false positives (<xref ref-type="bibr" rid="ref16">16</xref>). Indeed, as demonstrated by Bernstein et al. (<xref ref-type="bibr" rid="ref16">16</xref>), a faulty AI decreases radiographers&#x2019; accuracy especially if the results of the AI are shown in the final report.</p>
</sec>
<sec id="sec10">
<label>2.8</label>
<title>Satisfaction of search</title>
<p>&#x201C;The finding is missed because of the failure to continue to search for additional abnormalities after a first abnormality is found.&#x201D; This is a common situation, especially when advanced imaging modalities, such as CT or MRI, are evaluated (<xref ref-type="bibr" rid="ref10">10</xref>). To the best of the authors&#x2019; knowledge, no algorithm for lesion detection in advanced imaging modalities (CT or MRI) has been proposed in the veterinary literature, and, therefore, the usefulness of AI in the reduction of such an error has yet to be established.</p>
</sec>
<sec id="sec11">
<label>2.9</label>
<title>Complication</title>
<p>&#x201C;Complication from a procedure,&#x201D; meaning untoward events that could happen during an invasive examination procedure (<xref ref-type="bibr" rid="ref9">9</xref>). This is reported to be an uncommon type of error in human medicine (<xref ref-type="bibr" rid="ref10">10</xref>). The role of AI in the reduction of such an error is similar to that regarding other error types (e.g., prior examination).</p>
</sec>
<sec id="sec12">
<label>2.10</label>
<title>Satisfaction of report</title>
<p>&#x201C;The finding was missed because of the complacency of the report, and over-reliance on the radiology report of the previous examinations.&#x201D; This type of error arises from what is known as alliterative bias, meaning that one radiologist&#x2019;s judgement is influenced by that of another radiologist. To avoid this sort of bias (<xref ref-type="bibr" rid="ref6">6</xref>), suggest that the radiologist should read previous reports only after rendering the interpretation of findings. This is one of the most dangerous types of errors, as it is perpetuated from one study to the next (<xref ref-type="bibr" rid="ref10">10</xref>). The authors believe that AI could play a prominent role in reducing these error types. In fact, AI systems are unaware of the results of prior studies and could therefore help the clinician make more factual-based decisions that are devoid of cognitive biases.</p>
</sec>
</sec>
<sec sec-type="conclusions" id="sec13">
<label>3</label>
<title>Conclusion</title>
<p>A summary of the error types according to Kim and Mansfield (<xref ref-type="bibr" rid="ref10">10</xref>) and the possible contribution of AI-based products in their mitigation is reported in <xref ref-type="table" rid="tab1">Table 1</xref>. AI is still a very young technology in veterinary medicine and, despite the increasing number of applications available on the market, is far from being part of most practices&#x2019; clinical routine. The same is also true, to some extent, in human medicine. Indeed, despite the large investments in and the media impact of AI, the diffusion of AI-based systems is still limited, and actual improvements in healthcare quality related to the widespread adoption of these technologies are still to be demonstrated (<xref ref-type="bibr" rid="ref25">25</xref>).</p>
<table-wrap position="float" id="tab1">
<label>Table 1</label>
<caption><p>Possible errors according to Kim and Mansfield (<xref ref-type="bibr" rid="ref10">10</xref>) and role of AI in mitigation.</p></caption>
<table frame="hsides" rules="groups">
<thead>
<tr>
<th align="left" valign="top">Type of error</th>
<th align="left" valign="top">Role of AI</th>
</tr>
</thead>
<tbody>
<tr>
<td align="left" valign="bottom">Complacency</td>
<td align="left" valign="bottom">Yields lower number of false positives than radiologists do</td>
</tr>
<tr>
<td align="left" valign="top">Faulty reasoning</td>
<td align="left" valign="bottom">Limited usefulness of AI. Education plays a larger role in mitigating this error type</td>
</tr>
<tr>
<td align="left" valign="bottom">Lack of knowledge</td>
<td align="left" valign="bottom">Limited usefulness of AI (LLMs might be more effective)</td>
</tr>
<tr>
<td align="left" valign="bottom">Under-reading</td>
<td align="left" valign="bottom">Varies depending on the accuracy for each specific finding</td>
</tr>
<tr>
<td align="left" valign="top">Poor communication</td>
<td align="left" valign="bottom">Limited usefulness. LLMs could provide a means to homogenize the reports</td>
</tr>
<tr>
<td align="left" valign="top">Prior examination/history</td>
<td align="left" valign="bottom">AI-assisted reporting and AI-based tools to create quick summaries of the clinical history could help in mitigating this type of error</td>
</tr>
<tr>
<td align="left" valign="top">Location</td>
<td align="left" valign="bottom">AI scans the entire image/scan and is not influenced by the position of the lesion</td>
</tr>
<tr>
<td align="left" valign="top">Satisfaction of search</td>
<td align="left" valign="bottom">AI is unaware of the reasons for the scan/image and checks the entire exam</td>
</tr>
<tr>
<td align="left" valign="bottom">Complication</td>
<td align="left" valign="bottom">Similar to prior examination</td>
</tr>
<tr>
<td align="left" valign="top">Satisfaction of report</td>
<td align="left" valign="bottom">AI-based products are not influenced by this error type and could help in taking more factual-based decisions</td>
</tr>
</tbody>
</table>
<table-wrap-foot>
<p>LLMs, large language models.</p>
</table-wrap-foot>
</table-wrap>
<p>It is the authors&#x2019; opinion that AI will likely have different impacts on human and veterinary diagnostic imaging, mostly due to the intrinsic differences that exist between these two disciplines. The number of board-certified radiologists in veterinary medicine is still limited compared to those in human medicine, and therefore it is common practice for veterinary diagnostic images to be interpreted by non-specialists. This poses some questions regarding the effectiveness of these AI-based computer-aided systems in veterinary medicine. In fact, it is reported that AI has a variable accuracy for different radiographic findings (<xref ref-type="bibr" rid="ref18">18</xref>, <xref ref-type="bibr" rid="ref23">23</xref>). If the operator cannot determine the accuracy of the AI system&#x2019;s findings, relying on these systems might lead to misleading outcomes.</p>
<p>In the perspective article presented here, we did not address the importance of AI algorithms in assessing the quality of medical images. This application has been scarcely explored in veterinary medicine, and to date, only two studies highlights these algorithms as a promising tool to enhance the accuracy of interpreting canine radiographs by identifying technical errors (<xref ref-type="bibr" rid="ref26">26</xref>, <xref ref-type="bibr" rid="ref27">27</xref>). Conversely, in human medicine, numerous AI-based algorithms have been developed for evaluating the quality of chest X-ray images, showing promising results (<xref ref-type="bibr" rid="ref28">28</xref>, <xref ref-type="bibr" rid="ref29">29</xref>). This is a field where AI algorithms could again contribute to reducing radiologists&#x2019; interpretative error rates by automatically screening the quality of diagnostic images before interpretation, similar to what is already happening in human medicine.</p>
<p>In human medicine, new medical devices need to be approved by a regulatory agency, such as the European Medicines Agency in Europe or the Food and Drug Administration in the United States (<xref ref-type="bibr" rid="ref30">30</xref>). In veterinary medicine, such a regulatory agency does not exist and therefore, to date, there has not been a way to certify vendors&#x2019; claims regarding the accuracy and stability of the proposed systems (<xref ref-type="bibr" rid="ref31">31</xref>). It is the authors&#x2019; opinion that, in such a scenario in veterinary medicine, correct and impartial information to the final users is of vital importance in order to avoid misuse and possible fraud.</p>
</sec>
<sec sec-type="data-availability" id="sec14">
<title>Data availability statement</title>
<p>The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.</p>
</sec>
<sec sec-type="author-contributions" id="sec15">
<title>Author contributions</title>
<p>SB: Writing &#x2013; review &#x0026; editing. AZ: Writing &#x2013; review &#x0026; editing. TB: Writing &#x2013; original draft, Writing &#x2013; review &#x0026; editing.</p>
</sec>
</body>
<back>
<sec sec-type="funding-information" id="sec16">
<title>Funding</title>
<p>The author(s) declare that financial support was received for the research, authorship, and/or publication of this article. Open Access funding provided by Universit&#x00E0; degli Studi di Padova|University of Padua, Open Science Committee. The present paper was part of a project funded by a research grant from the Department of Animal Medicine, Production and Health &#x2013; MAPS, University of Padua, Italy: SID- Banzato 2023 (&#x20AC;20.000).</p>
</sec>
<sec sec-type="COI-statement" id="sec17">
<title>Conflict of interest</title>
<p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p>
<p>The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.</p>
</sec>
<sec sec-type="disclaimer" id="sec18">
<title>Publisher&#x2019;s note</title>
<p>All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.</p>
</sec>
<ref-list>
<title>References</title>
<ref id="ref1"><label>1.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Alexander</surname> <given-names>K</given-names></name></person-group>. <article-title>Reducing error in radiographic interpretation</article-title>. <source>Can Vet J</source>. (<year>2020</year>) <volume>51</volume>:<fpage>533</fpage>&#x2013;<lpage>536</lpage>.</citation></ref>
<ref id="ref2"><label>2.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Lamb</surname> <given-names>CR</given-names></name> <name><surname>Pfeiffer</surname> <given-names>DU</given-names></name> <name><surname>Mantis</surname> <given-names>P</given-names></name></person-group>. <article-title>Errors in radiographic interpretation made by veterinary students</article-title>. <source>J Vet Med Educ</source>. (<year>2007</year>) <volume>34</volume>:<fpage>157</fpage>&#x2013;<lpage>9</lpage>. doi: <pub-id pub-id-type="doi">10.3138/jvme.34.2.157</pub-id></citation></ref>
<ref id="ref3"><label>3.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Cohen</surname> <given-names>J</given-names></name> <name><surname>Fischetti</surname> <given-names>AJ</given-names></name> <name><surname>Daverio</surname> <given-names>H</given-names></name></person-group>. <article-title>Veterinary radiologic error rate as determined by necropsy</article-title>. <source>Vet Radiol Ultrasound</source>. (<year>2023</year>) <volume>64</volume>:<fpage>573</fpage>&#x2013;<lpage>84</lpage>. doi: <pub-id pub-id-type="doi">10.1111/vru.13259</pub-id></citation></ref>
<ref id="ref4"><label>4.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Waite</surname> <given-names>S</given-names></name> <name><surname>Kolla</surname> <given-names>S</given-names></name> <name><surname>Reede</surname> <given-names>D</given-names></name> <name><surname>Gale</surname> <given-names>B</given-names></name> <name><surname>Fuchs</surname> <given-names>T</given-names></name> <name><surname>Scott</surname> <given-names>J</given-names></name></person-group>. <article-title>Interpretive error in radiology</article-title>. <source>Am J Roentgenol</source>. (<year>2016</year>) <volume>208</volume>:<fpage>739</fpage>&#x2013;<lpage>49</lpage>. doi: <pub-id pub-id-type="doi">10.2214/ajr.16.16963</pub-id></citation></ref>
<ref id="ref5"><label>5.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Brady</surname> <given-names>AP</given-names></name></person-group>. <article-title>Error and discrepancy in radiology: inevitable or avoidable?</article-title> <source>Insights Imaging</source>. (<year>2017</year>) <volume>8</volume>:<fpage>171</fpage>&#x2013;<lpage>82</lpage>. doi: <pub-id pub-id-type="doi">10.1007/s13244-016-0534-1</pub-id></citation></ref>
<ref id="ref6"><label>6.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Itri</surname> <given-names>JN</given-names></name> <name><surname>Patel</surname> <given-names>SH</given-names></name></person-group>. <article-title>Heuristics and cognitive error in medical imaging</article-title>. <source>Am J Roentgenol</source>. (<year>2018</year>) <volume>210</volume>:<fpage>1097</fpage>&#x2013;<lpage>105</lpage>. doi: <pub-id pub-id-type="doi">10.2214/AJR.17.18907</pub-id></citation></ref>
<ref id="ref7"><label>7.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Itri</surname> <given-names>JN</given-names></name> <name><surname>Tappouni</surname> <given-names>RR</given-names></name> <name><surname>McEachern</surname> <given-names>RO</given-names></name> <name><surname>Pesch</surname> <given-names>AJ</given-names></name> <name><surname>Patel</surname> <given-names>SH</given-names></name></person-group>. <article-title>Fundamentals of diagnostic error in imaging</article-title>. <source>Radiographics</source>. (<year>2018</year>) <volume>38</volume>:<fpage>1845</fpage>&#x2013;<lpage>65</lpage>. doi: <pub-id pub-id-type="doi">10.1148/rg.2018180021</pub-id></citation></ref>
<ref id="ref8"><label>8.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Brook</surname> <given-names>OR</given-names></name> <name><surname>O&#x2019;Connell</surname> <given-names>AM</given-names></name> <name><surname>Thornton</surname> <given-names>E</given-names></name> <name><surname>Eisenberg</surname> <given-names>RL</given-names></name> <name><surname>Mendiratta-Lala</surname> <given-names>M</given-names></name> <name><surname>Kruskal</surname> <given-names>JB</given-names></name></person-group>. <article-title>Anatomy and pathophysiology of errors occurring in clinical radiology practice</article-title>. <source>Radiographics</source>. (<year>2010</year>) <volume>30</volume>:<fpage>1401</fpage>&#x2013;<lpage>10</lpage>. doi: <pub-id pub-id-type="doi">10.1148/rg.305105013</pub-id></citation></ref>
<ref id="ref9"><label>9.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Renfrew</surname> <given-names>DL</given-names></name> <name><surname>Franken</surname> <given-names>EA</given-names></name> <name><surname>Berbaum</surname> <given-names>KS</given-names></name> <name><surname>Weigelt</surname> <given-names>FH</given-names></name> <name><surname>Abu-Yousef</surname> <given-names>MM</given-names></name></person-group>. <article-title>Error in radiology: classification and lessons in 182 cases presented at a problem case conference</article-title>. <source>Radiology</source>. (<year>1992</year>) <volume>183</volume>:<fpage>145</fpage>&#x2013;<lpage>50</lpage>. doi: <pub-id pub-id-type="doi">10.1148/radiology.183.1.1549661</pub-id></citation></ref>
<ref id="ref10"><label>10.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Kim</surname> <given-names>YW</given-names></name> <name><surname>Mansfield</surname> <given-names>LT</given-names></name></person-group>. <article-title>Fool me twice: delayed diagnoses in radiology with emphasis on perpetuated errors</article-title>. <source>AJR Am J Roentgenol</source>. (<year>2014</year>) <volume>202</volume>:<fpage>465</fpage>&#x2013;<lpage>70</lpage>. doi: <pub-id pub-id-type="doi">10.2214/AJR.13.11493</pub-id></citation></ref>
<ref id="ref11"><label>11.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Bruno</surname> <given-names>MA</given-names></name> <name><surname>Walker</surname> <given-names>EA</given-names></name> <name><surname>Abujudeh</surname> <given-names>HH</given-names></name></person-group>. <article-title>Understanding and confronting our mistakes: the epidemiology of error in radiology and strategies for error reduction</article-title>. <source>Radiographics</source>. (<year>2015</year>) <volume>35</volume>:<fpage>1668</fpage>&#x2013;<lpage>76</lpage>. doi: <pub-id pub-id-type="doi">10.1148/rg.2015150023</pub-id></citation></ref>
<ref id="ref12"><label>12.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Hennessey</surname> <given-names>E</given-names></name> <name><surname>DiFazio</surname> <given-names>M</given-names></name> <name><surname>Hennessey</surname> <given-names>R</given-names></name> <name><surname>Cassel</surname> <given-names>N</given-names></name></person-group>. <article-title>Artificial intelligence in veterinary diagnostic imaging: a literature review</article-title>. <source>Vet Radiol Ultrasound</source>. (<year>2022</year>) <volume>63</volume>:<fpage>851</fpage>&#x2013;<lpage>70</lpage>. doi: <pub-id pub-id-type="doi">10.1111/vru.13163</pub-id></citation></ref>
<ref id="ref13"><label>13.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Basran</surname> <given-names>PS</given-names></name> <name><surname>Porter</surname> <given-names>I</given-names></name></person-group>. <article-title>Radiomics in veterinary medicine: overview, methods, and applications</article-title>. <source>Vet Radiol Ultrasound</source>. (<year>2022</year>) <volume>63</volume>:<fpage>828</fpage>&#x2013;<lpage>39</lpage>. doi: <pub-id pub-id-type="doi">10.1111/vru.13156</pub-id></citation></ref>
<ref id="ref14"><label>14.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Seah</surname> <given-names>JCY</given-names></name> <name><surname>Tang</surname> <given-names>CHM</given-names></name> <name><surname>Buchlak</surname> <given-names>QD</given-names></name> <name><surname>Holt</surname> <given-names>XG</given-names></name> <name><surname>Wardman</surname> <given-names>JB</given-names></name> <name><surname>Aimoldin</surname> <given-names>A</given-names></name> <etal/></person-group>. <article-title>Effect of a comprehensive deep-learning model on the accuracy of chest x-ray interpretation by radiologists: a retrospective, multireader multicase study</article-title>. <source>Lancet Digit Health</source>. (<year>2021</year>) <volume>3</volume>:<fpage>e496</fpage>&#x2013;<lpage>506</lpage>. doi: <pub-id pub-id-type="doi">10.1016/S2589-7500(21)00106-0</pub-id></citation></ref>
<ref id="ref15"><label>15.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Boissady</surname> <given-names>E</given-names></name> <name><surname>de La Comble</surname> <given-names>A</given-names></name> <name><surname>Zhu</surname> <given-names>X</given-names></name> <name><surname>Hespel</surname> <given-names>AM</given-names></name></person-group>. <article-title>Artificial intelligence evaluating primary thoracic lesions has an overall lower error rate compared to veterinarians or veterinarians in conjunction with the artificial intelligence</article-title>. <source>Vet Radiol Ultrasound</source>. (<year>2020</year>) <volume>61</volume>:<fpage>619</fpage>&#x2013;<lpage>27</lpage>. doi: <pub-id pub-id-type="doi">10.1111/vru.12912</pub-id></citation></ref>
<ref id="ref16"><label>16.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Bernstein</surname> <given-names>MH</given-names></name> <name><surname>Atalay</surname> <given-names>MK</given-names></name> <name><surname>Dibble</surname> <given-names>EH</given-names></name> <name><surname>Maxwell</surname> <given-names>AWP</given-names></name> <name><surname>Karam</surname> <given-names>AR</given-names></name> <name><surname>Agarwal</surname> <given-names>S</given-names></name> <etal/></person-group>. <article-title>Can incorrect artificial intelligence (AI) results impact radiologists, and if so, what can we do about it? A multi-reader pilot study of lung cancer detection with chest radiography</article-title>. <source>Eur Radiol</source>. (<year>2023</year>) <volume>33</volume>:<fpage>8263</fpage>&#x2013;<lpage>9</lpage>. doi: <pub-id pub-id-type="doi">10.1007/s00330-023-09747-1</pub-id></citation></ref>
<ref id="ref17"><label>17.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Degnan</surname> <given-names>AJ</given-names></name> <name><surname>Ghobadi</surname> <given-names>EH</given-names></name> <name><surname>Hardy</surname> <given-names>P</given-names></name> <name><surname>Krupinski</surname> <given-names>E</given-names></name> <name><surname>Scali</surname> <given-names>EP</given-names></name> <name><surname>Stratchko</surname> <given-names>L</given-names></name> <etal/></person-group>. <article-title>Perceptual and interpretive error in diagnostic radiology&#x2014;causes and potential solutions</article-title>. <source>Acad Radiol</source>. (<year>2019</year>) <volume>26</volume>:<fpage>833</fpage>&#x2013;<lpage>45</lpage>. doi: <pub-id pub-id-type="doi">10.1016/j.acra.2018.11.006</pub-id></citation></ref>
<ref id="ref18"><label>18.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Banzato</surname> <given-names>T</given-names></name> <name><surname>Wodzinski</surname> <given-names>M</given-names></name> <name><surname>Burti</surname> <given-names>S</given-names></name> <name><surname>Osti</surname> <given-names>VL</given-names></name> <name><surname>Rossoni</surname> <given-names>V</given-names></name> <name><surname>Atzori</surname> <given-names>M</given-names></name> <etal/></person-group>. <article-title>Automatic classification of canine thoracic radiographs using deep learning</article-title>. <source>Sci Rep</source>. (<year>2021</year>) <volume>11</volume>:<fpage>3964</fpage>. doi: <pub-id pub-id-type="doi">10.1038/s41598-021-83515-3</pub-id></citation></ref>
<ref id="ref19"><label>19.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Banzato</surname> <given-names>T</given-names></name> <name><surname>Wodzinski</surname> <given-names>M</given-names></name> <name><surname>Tauceri</surname> <given-names>F</given-names></name> <name><surname>Don&#x00E0;</surname> <given-names>C</given-names></name> <name><surname>Scavazza</surname> <given-names>F</given-names></name> <name><surname>M&#x00FC;ller</surname> <given-names>H</given-names></name> <etal/></person-group>. <article-title>An AI-based algorithm for the automatic classification of thoracic radiographs in cats</article-title>. <source>Front Vet Sci</source>. (<year>2021</year>) <volume>8</volume>:<fpage>731936</fpage>. doi: <pub-id pub-id-type="doi">10.3389/fvets.2021.731936</pub-id></citation></ref>
<ref id="ref20"><label>20.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Mese</surname> <given-names>I</given-names></name> <name><surname>Taslicay</surname> <given-names>CA</given-names></name> <name><surname>Sivrioglu</surname> <given-names>AK</given-names></name></person-group>. <article-title>Improving radiology workflow using ChatGPT and artificial intelligence</article-title>. <source>Clin Imaging</source>. (<year>2023</year>) <volume>103</volume>:<fpage>109993</fpage>. doi: <pub-id pub-id-type="doi">10.1016/j.clinimag.2023.109993</pub-id></citation></ref>
<ref id="ref21"><label>21.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Bradshaw</surname> <given-names>TJ</given-names></name> <name><surname>Huemann</surname> <given-names>Z</given-names></name> <name><surname>Hu</surname> <given-names>J</given-names></name> <name><surname>Rahmim</surname> <given-names>A</given-names></name></person-group>. <article-title>A guide to cross-validation for artificial intelligence in medical imaging</article-title>. <source>Radiol Artif Intell</source>. (<year>2023</year>) <volume>5</volume>:<fpage>e220232</fpage>. doi: <pub-id pub-id-type="doi">10.1148/ryai.220232</pub-id></citation></ref>
<ref id="ref22"><label>22.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>M&#x00FC;ller</surname> <given-names>TR</given-names></name> <name><surname>Solano</surname> <given-names>M</given-names></name> <name><surname>Tsunemi</surname> <given-names>MH</given-names></name></person-group>. <article-title>Accuracy of artificial intelligence software for the detection of confirmed pleural effusion in thoracic radiographs in dogs</article-title>. <source>Vet Radiol Ultrasound</source>. (<year>2022</year>) <volume>63</volume>:<fpage>573</fpage>&#x2013;<lpage>9</lpage>. doi: <pub-id pub-id-type="doi">10.1111/vru.13089</pub-id></citation></ref>
<ref id="ref23"><label>23.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Pomerantz</surname> <given-names>LK</given-names></name> <name><surname>Solano</surname> <given-names>M</given-names></name> <name><surname>Kalosa-Kenyon</surname> <given-names>E</given-names></name> <name><surname>Leah</surname> <given-names>C</given-names></name> <name><surname>Pomerantz</surname> <given-names>K</given-names></name></person-group>. <article-title>Performance of a commercially available artificial intelligence software for the detection of confirmed pulmonary nodules and masses in canine thoracic radiography</article-title>. <source>Vet Radiol Ultrasound</source>. (<year>2023</year>) <volume>64</volume>:<fpage>881</fpage>&#x2013;<lpage>9</lpage>. doi: <pub-id pub-id-type="doi">10.1111/vru.13287</pub-id></citation></ref>
<ref id="ref24"><label>24.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Drew</surname> <given-names>T</given-names></name> <name><surname>Evans</surname> <given-names>K</given-names></name> <name><surname>V&#x00F5;</surname> <given-names>MLH</given-names></name> <name><surname>Jacobson</surname> <given-names>FL</given-names></name> <name><surname>Wolfe</surname> <given-names>JM</given-names></name></person-group>. <article-title>Informatics in radiology: what can you see in a single glance and how might this guide visual search in medical images?</article-title> <source>Radiographics</source>. (<year>2013</year>) <volume>33</volume>:<fpage>263</fpage>&#x2013;<lpage>74</lpage>. doi: <pub-id pub-id-type="doi">10.1148/rg.331125023</pub-id></citation></ref>
<ref id="ref25"><label>25.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>van Leeuwen</surname> <given-names>KG</given-names></name> <name><surname>de Rooij</surname> <given-names>M</given-names></name> <name><surname>Schalekamp</surname> <given-names>S</given-names></name> <name><surname>van Ginneken</surname> <given-names>B</given-names></name> <name><surname>Rutten</surname> <given-names>MJCM</given-names></name></person-group>. <article-title>How does artificial intelligence in radiology improve efficiency and health outcomes?</article-title> <source>Pediatr Radiol</source>. (<year>2022</year>) <volume>52</volume>:<fpage>2087</fpage>&#x2013;<lpage>93</lpage>. doi: <pub-id pub-id-type="doi">10.1007/s00247-021-05114-8</pub-id></citation></ref>
<ref id="ref26"><label>26.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Banzato</surname> <given-names>T</given-names></name> <name><surname>Wodzinski</surname> <given-names>M</given-names></name> <name><surname>Burti</surname> <given-names>S</given-names></name> <name><surname>Vettore</surname> <given-names>E</given-names></name> <name><surname>Muller</surname> <given-names>H</given-names></name> <name><surname>Zotti</surname> <given-names>A</given-names></name></person-group>. <article-title>An AI-based algorithm for the automatic evaluation of image quality in canine thoracic radiographs</article-title>. <source>Sci Rep</source>. (<year>2023</year>) <volume>13</volume>:<fpage>17024</fpage>. doi: <pub-id pub-id-type="doi">10.1038/s41598-023-44089-4</pub-id></citation></ref>
<ref id="ref27"><label>27.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Tahghighi</surname> <given-names>P</given-names></name> <name><surname>Norena</surname> <given-names>N</given-names></name> <name><surname>Ukwatta</surname> <given-names>E</given-names></name> <name><surname>Appleby</surname> <given-names>RB</given-names></name> <name><surname>Komeili</surname> <given-names>A</given-names></name></person-group>. <article-title>Automatic classification of symmetry of hemithoraces in canine and feline radiographs</article-title>. <source>J Med Imaging</source>. (<year>2023</year>) <volume>10</volume>:<fpage>044004</fpage>. doi: <pub-id pub-id-type="doi">10.1117/1.JMI.10.4.044004</pub-id></citation></ref>
<ref id="ref28"><label>28.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Meng</surname> <given-names>Y</given-names></name> <name><surname>Ruan</surname> <given-names>J</given-names></name> <name><surname>Yang</surname> <given-names>B</given-names></name> <name><surname>Gao</surname> <given-names>Y</given-names></name> <name><surname>Jin</surname> <given-names>J</given-names></name> <name><surname>Dong</surname> <given-names>F</given-names></name> <etal/></person-group>. <article-title>Automated quality assessment of chest radiographs based on deep learning and linear regression cascade algorithms</article-title>. <source>Eur Radiol</source>. (<year>2022</year>):<fpage>7680</fpage>&#x2013;<lpage>90</lpage>. doi: <pub-id pub-id-type="doi">10.1007/s00330-022-08771-x</pub-id></citation></ref>
<ref id="ref29"><label>29.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Nousiainen</surname> <given-names>K</given-names></name> <name><surname>M&#x00E4;kel&#x00E4;</surname> <given-names>T</given-names></name> <name><surname>Piilonen</surname> <given-names>A</given-names></name> <name><surname>Peltonen</surname> <given-names>JI</given-names></name></person-group>. <article-title>Automating chest radiograph imaging quality control</article-title>. <source>Phys Med</source>. (<year>2021</year>) <volume>83</volume>:<fpage>138</fpage>&#x2013;<lpage>45</lpage>. doi: <pub-id pub-id-type="doi">10.1016/j.ejmp.2021.03.014</pub-id></citation></ref>
<ref id="ref30"><label>30.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Darrow</surname> <given-names>JJ</given-names></name> <name><surname>Avorn</surname> <given-names>J</given-names></name> <name><surname>Kesselheim</surname> <given-names>AS</given-names></name></person-group>. <article-title>FDA regulation and approval of medical devices: 1976&#x2013;2020</article-title>. <source>JAMA</source>. (<year>2021</year>) <volume>326</volume>:<fpage>420</fpage>&#x2013;<lpage>32</lpage>. doi: <pub-id pub-id-type="doi">10.1001/jama.2021.11171</pub-id></citation></ref>
<ref id="ref31"><label>31.</label><citation citation-type="journal"><person-group person-group-type="author"><name><surname>Cohen</surname> <given-names>EB</given-names></name> <name><surname>Gordon</surname> <given-names>IK</given-names></name></person-group>. <article-title>First, do no harm. Ethical and legal issues of artificial intelligence and machine learning in veterinary radiology and radiation oncology</article-title>. <source>Vet Radiol Ultrasound</source>. (<year>2022</year>) <volume>63</volume>:<fpage>840</fpage>&#x2013;<lpage>50</lpage>. doi: <pub-id pub-id-type="doi">10.1111/vru.13171</pub-id></citation></ref>
</ref-list>
</back>
</article>
