Automating Study Report and
SEND Integration in
Nonclinical Safety Studies

Yes, Single-Track can be implemented at any point. Ideally, to maximize the benefits of Single-Track, it is best to implement as soon as the protocol is signed. For CROs with multiple studies, the benefits are realized over time.

Yes, reusing templates across studies is a core design principle. Templates are managed centrally in Xbiom's MDR (Metadata Repository), which is a system of record. Templates are organized by Study Report Type (e.g., repeat-dose, single-dose) within the MDR.

• Organized by study type: templates are structured by Study Report Type so the right template is always matched to the right study design.
• Consistent structure, study-specific content: the same template drives generation across all studies of that type; the AI populates it with each study's unique data while the format and sections remain standardized.
• Set up once, reused continuously: templates are configured during the implementation and immediately available for all subsequent studies, with the Study Director able to view the live template alongside generated content at any time.

Sub-reports are accepted primarily as PDFs, whether LIMS-generated, externally sourced lab reports, or even scanned handwritten pages. Certificates of Analysis (test article, vehicle, diet/feed, bedding, and water) and formulation/analytical records in standard document formats (DOCX & Excel) are also supported. Any additional formats required as appendices or addenda to the study report are accommodated as needed.

Xbiom's Single Track process automates SEND generation through several layers of logic and guided workflows:

• Trial Design automation — the Trial Design Editor automatically populates TS, TE, TA, and TX domains directly from the protocol document, with SEND team review propagating decisions to all downstream domains.
• Auto-derivation of SEND variables — derived fields such as –DY, –NOMDY, –BLFL, and TPT variables are automatically generated from available date-time stamps.
• ORRES splitting and CT normalization — findings are automatically split into original finding, specimen, and laterality, with controlled terminology mappings applied and confirmed or overridden by the SEND team.
• Unit Conversion & Normalization — automatically converts source measurement units to CDISC standard units (ORRESU, STRESU).
• MDR propagation — methods, instruments, and practices captured in the Metadata Repository are automatically propagated across relevant SEND domains.
• Supplemental Domain Generation — SUPP datasets are automatically generated based on the IG selected and data mappings.
• Draft SEND preview early — the SEND QC team can generate and review draft XPTs at any time before the Final Report, surfacing gaps and conformance issues well before submission.

Yes — validation against FDA business/validation rules and CDISC compliance rules is fully supported through Xbiom's eDataValidator (eDV), built directly into the SingleTrack platform:

Errors and warnings in the QC dashboard are interactive — each flag links directly to where the issue occurs in the dataset. The SEND QC team can run these checks at any time during draft SEND generation, making validation an ongoing process rather than a last-minute gate before submission.

SingleTrack is designed to accommodate organizational and Study Director style preferences at multiple levels:

1. Templates built around your organization's style. Your existing report templates and Tables of Contents are loaded directly into the MDR at setup. The AI generates all report sections using your templates as the formatting and structural baseline. LLM prompts are also configured specifically for your CRO's style preferences from the outset.

2. Study Director controls the final text — always. The AI is an assistant, not the author. The Study Director controls the final published text through three mechanisms:

• Direct editing — any AI-generated section can be edited directly in the interface.
• Prompt-driven regeneration — the Study Director can provide natural language inputs to adjust tone, style, or format (e.g., “re-draft with softer language” or “reformat findings as a table”) and the AI regenerates the section while preserving data integrity.
• Style reference — the Study Director can reference an existing sample report section and the AI will adopt that style across the regenerated output.

Yes — waiting for histopathology does not mean waiting to start. Sections covering study conduct, methods, test item, formulations, clinical pathology, in-life observations, and other completed domains can all be drafted and reviewed while histopathology is still pending. Your report could be ready the same day you receive histopathology results.

Each report section moves through a clear three-stage workflow — Draft → Ready for Review → Completed — independently of other sections. The system provides a real-time view of the report's completeness, making clear which sections are finalized and which are awaiting data.

Once histopathology data is received, it is ingested into Xbiom and the AI generates the histopathology results and summary sections through the same workflow. Draft Conclusions and Summary sections are generated last, after all domain-level results are in place.

If you have an in-house SEND team — the system makes them significantly more efficient. SingleTrack automates the heavy lifting of SEND generation so your existing team shifts from manual, repetitive data work to a review and confirmation role. The same team can handle a much larger study volume without adding headcount.

If you do not have an in-house SEND team — the system is designed to enable that too. SingleTrack's automation means SEND generation no longer requires a large team of SDTM/SEND specialists. The Transformer automatically produces submission-ready SEND XPT files directly from the locked data layer without manual intervention.

PointCross also provides ongoing support — including weekly office hours, on-demand assistance, and support for novel or complex study designs while your internal capability develops.

Data accuracy is addressed at multiple layers, starting from the source rather than being checked after the fact:

• Data flows directly from source — no manual transcription. LIMS data is ingested automatically; CoAs, sub-reports, and facility records are ingested directly into Xbiom. The AI generates report text from this integrated data layer, eliminating the primary source of transcription error.
• Protocol vs. actual automated comparison. The system automatically compares protocol-defined details against actual study conduct data, generating alerts and flagging deviations for reporting.
• AI-generated text validated against source data. LLMs interact with the platform exclusively through read-only interfaces and cannot modify source data. The system automatically validates all AI-generated text to ensure no material change in meaning relative to the data-supported findings.
• Full source traceability visible to the Study Director. Every generated report section shows where each piece of content came from — template, protocol, CoA, LIMS data — side by side with the generated text.
• Human-in-the-loop review at every stage. All AI-generated content is subject to Study Director review, edit, and sign-off. The system does not publish anything autonomously.
• 21 CFR Part 11 controls. All modifications are governed by 21 CFR Part 11 electronic records controls, providing a full audit trail of every change made to report content.

Yes — SingleTrack provides full bi-directional data traceability. Every data point, signal, and finding in the study report links directly to its source in as-collected data, and conversely, every raw data point can be traced forward to where and how it appears in the final report and SEND dataset. This traceability is visible to the Study Director at all times and is critical for audit readiness and FDA/PMDA regulatory submission quality.

Every amendment is captured and extracted automatically. When an amendment is uploaded, the AI identifies and summarizes the key changes, with each change traceable back to its source in the amendment document.

The study schedule and design always reflect the current protocol version. Planned vs. actual data collections are continuously tracked against the active protocol version, with deviations flagged and annotated automatically in real time — building the deviation log as the study progresses rather than reconstructing it at the end.

Amendments propagate through to SEND. Protocol changes are automatically reflected in the SEND trial design domains, keeping SEND metadata in sync with the current approved version of the protocol.

Post-submission amendments are versioned too. If a report or SEND dataset needs to be amended after final publication, the system supports a formal re-lock process producing a new versioned deliverable with a full audit trail.