<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liu-etal-2026-routeverify">
    <titleInfo>
        <title>RouteVerify: A Diagnostic Study of Tool-Augmented Chest X-ray Agents</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Haochen</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yukuo</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xia</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Qingwei</namePart>
        <namePart type="family">Wang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Shourong</namePart>
        <namePart type="family">Xin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zhuohui</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Long</namePart>
        <namePart type="family">Zhao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-462-0</identifier>
    </relatedItem>
    <abstract>Tool-augmented medical agents promise clinical control and interpretability, yet their specialized outputs are often invoked broadly and compressed into text. It remains unclear whether failures arise from this textual interface or low-fidelity upstream tools. We study this question through RouteVerify, a transparent diagnostic sandbox combining a lightweight linear router, compact evidence cards, and verifier-gated multimodal reinspection. On matched and held-out ChestAgentBench splits, RouteVerify improves answer accuracy by about seven points over naive always-call-tools execution while roughly halving token use. However, it remains about two points below a direct VLM-only baseline and uses nearly twice the tokens. Oracle probes clarify the bottleneck: replacing predicted evidence with benchmark-side gold explanations raises the same downstream text reasoner above 95 percent accuracy, or above 91 percent after removing explicit answer cues. Thus, text cards can express sufficient evidence on this benchmark; the dominant practical limitation is upstream tool quality. An external JSRT/SCR audit supports this diagnosis, revealing limited nodule-classification discrimination and structure-dependent segmentation quality. The remaining raw-versus-sanitized oracle gap also indicates sensitivity to human-authored cues. We present RouteVerify as an empirical boundary study of where current tool-augmented medical agents fail.</abstract>
    <identifier type="citekey">liu-etal-2026-routeverify</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.813/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>17213</start>
            <end>17225</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T RouteVerify: A Diagnostic Study of Tool-Augmented Chest X-ray Agents
%A Liu, Haochen
%A Li, Yukuo
%A Liu, Xia
%A Wang, Qingwei
%A Xin, Shourong
%A Liu, Zhuohui
%A Zhao, Long
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F liu-etal-2026-routeverify
%X Tool-augmented medical agents promise clinical control and interpretability, yet their specialized outputs are often invoked broadly and compressed into text. It remains unclear whether failures arise from this textual interface or low-fidelity upstream tools. We study this question through RouteVerify, a transparent diagnostic sandbox combining a lightweight linear router, compact evidence cards, and verifier-gated multimodal reinspection. On matched and held-out ChestAgentBench splits, RouteVerify improves answer accuracy by about seven points over naive always-call-tools execution while roughly halving token use. However, it remains about two points below a direct VLM-only baseline and uses nearly twice the tokens. Oracle probes clarify the bottleneck: replacing predicted evidence with benchmark-side gold explanations raises the same downstream text reasoner above 95 percent accuracy, or above 91 percent after removing explicit answer cues. Thus, text cards can express sufficient evidence on this benchmark; the dominant practical limitation is upstream tool quality. An external JSRT/SCR audit supports this diagnosis, revealing limited nodule-classification discrimination and structure-dependent segmentation quality. The remaining raw-versus-sanitized oracle gap also indicates sensitivity to human-authored cues. We present RouteVerify as an empirical boundary study of where current tool-augmented medical agents fail.
%U https://aclanthology.org/2026.findings-emnlp.813/
%P 17213-17225

Markdown (Informal)

[RouteVerify: A Diagnostic Study of Tool-Augmented Chest X-ray Agents](https://aclanthology.org/2026.findings-emnlp.813/) (Liu et al., Findings 2026)

• RouteVerify: A Diagnostic Study of Tool-Augmented Chest X-ray Agents (Liu et al., Findings 2026)

ACL

• Haochen Liu, Yukuo Li, Xia Liu, Qingwei Wang, Shourong Xin, Zhuohui Liu, and Long Zhao. 2026. RouteVerify: A Diagnostic Study of Tool-Augmented Chest X-ray Agents. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 17213–17225, Budapest, Hungary. Association for Computational Linguistics.