Incidental
pulmonary nodules are now a routine part of clinical practice. With the
expansion of CT imaging, IPN programs and lung cancer screening, detection has
become more consistent than ever before.

What
remains difficult is everything that follows.

Managing
nodules requires more than identifying them. It involves understanding risk,
defining follow-up, tracking patients over time, and ensuring that recommended
care is completed. As volumes increase, this becomes difficult to sustain
through manual workflows alone.

Recent advances in technology are
beginning to address this problem, not by changing how clinicians make
decisions, but by improving how consistently those decisions are executed.

The technologies shaping
nodule management today

Current solutions used in pulmonary
nodule management and closed-loop follow-up can be broadly grouped into five
categories:

Each of these addresses a different gap
in the care pathway.

The first wave of innovation in this
space focused on improving detection.

AI-based
detection systems analyze imaging data, primarily CT scans, to identify
pulmonary nodules with high sensitivity. These systems are particularly helpful
for small or subtle nodules that may be difficult to detect consistently,
especially in high-volume settings.

However,
in real-world practice, detection remains an ongoing challenge, particularly
outside structured screening programs.

A
large proportion of patients first enter the system through <a href="https://www.qure.ai/blog/role-of-an-ai-for-reliable-screening-of-abnormality-in-x-rays">routine chest X-rays</a>, where nodules can be subtle,
overlapping with anatomical structures, or deprioritized in busy workflows.
This is where variability is highest, and where clinically meaningful findings
are most likely to be missed.

AI-based detection on chest X-rays is
beginning to address this gap.

For
example, solutions such as qXR act as a second reader on routine imaging,
highlighting suspected nodules in the 6–30 mm range and bringing attention to
findings that may otherwise go unnoticed. In real-world settings, this has
translated into the ability to identify up to 94% of nodules that were
initially missed, while operating at a sensitivity of 83%,
particularly for subtle findings in everyday clinical workflows1.

This
is particularly relevant because chest X-rays are often the first imaging
touchpoint, not just in screening but across emergency, outpatient, and
inpatient settings. Improving detection at this stage effectively expands the
front end of the pathway, allowing more patients to enter appropriate follow-up
earlier.

At
scale, this approach has already been implemented widely, with imaging AI
systems deployed across 5,200+ sites and contributing to impacting over 43
million lives, reflecting both clinical adoption and real-world utility.

Detection
technologies, therefore, are not just about improving sensitivity in controlled
environments. They are about reducing variability in routine care, where
the majority of patients are first seen.

Volumetric measurement and
longitudinal analysis tools

Once a nodule is detected, the focus
shifts to understanding its behavior over time.

Traditional
measurement approaches rely on diameter, which can vary depending on technique
and interpretation. Small differences in measurement can lead to uncertainty,
particularly when assessing growth.

Volumetric analysis offers a more precise
alternative.

By
evaluating nodules in three dimensions, these tools allow clinicians to track
changes in volume across serial scans. This is clinically important because
growth is one of the strongest indicators of malignancy.

In
practice, volumetric tools reduce inter-reader variability and provide a more
reliable basis for comparing nodules over time. They also make it easier to
detect subtle changes that may not be apparent through diameter measurements
alone.

This shift from single timepoint
assessment to longitudinal analysis is central to modern nodule management.

In
real-world settings, this is increasingly supported by systems that combine quantitative
measurement with longitudinal tracking. <a href="https://www.qure.ai/us">AI healthcare solutions</a> such as qCT enable semi-automated
nodule segmentation and provide standardized measurements across timepoints,
including diameter, volume, and density. This allows clinicians to move beyond
visual comparison and rely on reproducible metrics when assessing change.

An important addition is the ability to
calculate volume doubling time (VDT) across serial scans, which provides
a more objective indicator of growth dynamics. When combined with integrated
clinical frameworks such as Lung-RADS classification, Brock risk scoring, and
guideline-based follow-up recommendations, this creates a more structured
approach to decision-making.

What this enables in practice is a shift
from:

This
level of precision becomes particularly important in cases where growth is
subtle, and where small differences in interpretation can significantly
influence management decisions.

Structured reporting and
guideline integration systems

Another important area of advancement is how findings are documented and communicated.

<a href="https://www.qure.ai/blog/teaching-machines-to-read-radiology-reports">Radiology
reports</a> have traditionally been narrative, which can lead to
variability in how nodules are described and how follow-up recommendations are
conveyed.

Structured reporting systems aim to
address this by integrating guideline-based frameworks, such as Lung-RADS or
Fleischner, directly into reporting workflows.

Instead of relying solely on descriptive
text, these systems provide:

This improves communication between
radiologists and referring clinicians and reduces ambiguity in management
decisions.

In
practice, structured reporting helps ensure that recommendations are not only
made, but are also actionable.

Report interpretation: from
NLP to LLM-based systems

Even
with improvements in reporting, a large amount of clinically relevant
information remains embedded in unstructured text.

Natural
language processing (NLP) has been used to extract information from radiology
reports by identifying keywords and predefined patterns. These systems can
detect mentions of nodules and capture basic attributes such as size or
follow-up recommendations.

However, NLP approaches often struggle
with variability in language and lack the ability to fully interpret clinical
context. They may miss nuanced findings or incorrectly flag non-actionable
mentions.

More recent approaches using large
language models (LLMs) are beginning to address these limitations.

Unlike traditional NLP, LLM-based systems
are designed to interpret meaning rather than simply extract terms. This allows
them to:

In
practical terms, this reduces the need for manual chart review and improves the
accuracy of identifying patients who require follow-up.

The
shift from NLP to LLM-based interpretation represents a move from pattern
recognition to contextual understanding, which is particularly important in
complex clinical narratives.

Longitudinal tracking and
closed-loop workflow platforms

The most critical gap in nodule management has historically been follow-up over time.

Electronic medical records are effective
at documenting care, but they are not designed to track patients longitudinally
in a structured way. This creates a disconnect between recommendations and
execution.

<a href="https://www.qure.ai/blog/how-lung-nodule-tracking-software-supports-nurse-navigators">Longitudinal
tracking systems</a> are designed specifically to address this.

These
platforms maintain a continuous view of patients with nodules, tracking:

More
importantly, they introduce a time-based workflow, where patients are surfaced
when action is required and missed follow-ups are flagged.

This transforms follow-up from a passive
process into an actively managed one.

Studies
have shown that a substantial proportion of patients with pulmonary nodules do
not receive appropriate follow-up. Closed-loop systems are designed to reduce
this gap by ensuring that patients remain within the pathway until follow-up is
completed.

In
practice, this is where dedicated tracking layers become essential. Platforms
such as qTrack are designed to manage patients longitudinally across both
screening and incidental findings, creating a single, unified view of
follow-up. By bringing together imaging data, recommendations, real-time
customizable dashboards and follow-up status into one workflow, these systems
help ensure that patients do not fall out of the pathway between scans.

What
this enables is not just tracking, but active coordination, where
patients who require action are surfaced, prioritized, and managed in a
structured way. For clinical teams, this reduces reliance on manual tracking
and fragmented workflows, allowing follow-up to be handled more consistently at
scale.

Individually, each of these technologies
addresses a specific part of the problem.

The
real value emerges when these layers are integrated.

Closed-loop systems bring these
components together into a continuous workflow, where findings move from
detection to follow-up without being lost in transition.

As
imaging volumes continue to grow, maintaining this level of coordination
manually becomes increasingly difficult.

qTrack
is designed to support this infrastructure layer by providing longitudinal
tracking across both screening and incidental findings. By maintaining
visibility across imaging and time, and organizing patients within a structured
workflow, these systems help ensure that recommended follow-up is completed.

This allows clinical teams to focus on
interpretation and decision-making, while reducing gaps in execution.

The management of incidental pulmonary
nodules is evolving.

The next phase of progress is not about
detecting more nodules.
&nbsp;It is about managing them more reliably.

Advances
in technology are enabling this shift by improving consistency, reducing
variability, and ensuring that patients remain visible within the care pathway.

Ultimately,
the goal is straightforward: every detected nodule should lead to appropriate,
timely follow-up.

Schedule a consultation
with an expert today.&nbsp;Reach out to us on partner@qure.ai. 

large_4blog_thumbnail.webp

medium_4blog_thumbnail.webp

small_4blog_thumbnail.webp

thumbnail_4blog_thumbnail.webp

4blog_thumbnail.webp

What Are the Latest Technologies Used in the Management and Closed-Loop Follow-Up of Incidental Pulmonary Nodules?

In routine clinical practice, incidental
lung nodules are not rare findings. They are encountered across a wide range of
imaging studies - CT scans performed for trauma, cardiac evaluation, or
non-respiratory complaints often reveal unexpected abnormalities in the lung.

Pulmonary nodules are detected in up
to 25-50% of <a href="https://www.qure.ai/blog/qct-lung-catching-lung-cancer-early">chest CT scans</a>,
depending on the patient population and imaging protocols1. For most
clinicians, the detection itself is not the challenge. The real issue begins
after the finding is documented.

In many cases, a nodule is noted in the
radiology report, a follow-up is suggested, and then the patient disappears
from the pathway. Months later, that same patient may reappear with disease
progression that could have been addressed earlier.

This is not a failure of individual
clinicians. It reflects how our systems are currently designed.

Missed follow-up is rarely due to a
single breakdown. It is the result of multiple small gaps across the care
pathway.

1.
Findings Are Documented but Not Acted Upon

Radiology reports often include follow-up
recommendations, but these are embedded within narrative text and not always
translated into a clear next step.

In practice, what is missing is not
information, but operational clarity:

Without these elements, even
well-documented findings may not lead to action.

2.EMR Systems Do Not Support Longitudinal Tracking

Electronic medical records are designed
to store information, but they are not built to manage longitudinal <a href="https://www.qure.ai/blog/on-deep-learning-healthcare-and-qure-ai">clinical
workflows</a>.

This creates a gap between what is
recommended and what actually happens.

3.The Handoff Between Specialties Is Variable

The transition from radiology to
pulmonology or primary care is a critical step, and often an inconsistent one.

A recommendation such as “consider
follow-up” leaves room for interpretation. Without a defined interval, clear
ownership, or a structured pathway, the next step may be delayed or missed
entirely.

In many cases, responsibility for
follow-up is not clearly assigned.

Depending on the setting, this may fall
to:

High-functioning
programs address this by assigning follow-up to a defined pathway,
rather than an individual clinician. This is often operationalized through a
lung nodule or IPN program with centralized oversight.

4.Follow-Up Relies on Memory Instead of Systems

In the absence of structured tracking,
follow-up often depends on:

This approach does not scale,
particularly in systems managing large imaging volumes. Studies suggest that up
to 50% of patients with incidental pulmonary nodules do not receive appropriate
follow-up2.

A reliable system
does not depend on remembering, it depends on visibility and tracking.

What Actually Works:
Building a Reliable Follow-Up System

Improving follow-up requires a shift from
reactive workflows to structured, longitudinal management. It requires systems
designed for reliability.

Effective follow-up begins with clarity.
Each finding should be associated with:

The difference between “consider
follow-up” and “repeat CT in 6 months” is operationally significant.

2.Treat Follow-Up
as a Longitudinal Workflow

Follow-up is not a single event. It is a
process that extends across time and encounters.

This requires moving
from episodic care to programmatic tracking.

3.Centralize
Visibility Across the Program

In high-performing programs, follow-up is
not managed through individual inboxes.

This allows teams to focus on patients
who need action, rather than searching through reports.

Many systems rely on nurse navigators or
coordinators to manage follow-up.

When these roles are supported by
structured systems, they can manage significantly larger patient volumes with
greater consistency.

A closed-loop system is not defined by
documentation. It is defined by completion.

Anything short of this leaves the pathway
open.

The Role of Technology
in Supporting Reliability

Technology is most effective when it
reduces dependence on manual processes.

When <a href="https://www.qure.ai/blog/what-we-learned-deploying-deep-learning-at-scale-for-radiology-images">integrated into
existing workflows</a>, this shifts follow-up from a reactive process to
a managed system.

Why This
Matters for Patient Outcomes

The goal of managing lung nodules is not
simply documentation. It is timely intervention when needed.

Incidental lung nodules are a routine
part of clinical practice. The challenge is not in detection, but in ensuring
that findings lead to appropriate action.

Missed follow-up reflects a gap between clinical
intent and system capability.

In many health systems, this is enabled
through infrastructure that supports tracking across both <a href="https://www.qure.ai/blog/scaling-up-tb-screening-with-ai-deploying-automated-x-ray-screening-in-remote-regions">screening
populations</a> and incidental findings.

Platforms such as qTrack function as a
longitudinal layer within the care pathway, helping maintain visibility of
patients over time and supporting teams in ensuring that recommended follow-up
is completed.

By organizing patients within a
structured workflow and reducing reliance on manual tracking, this approach
allows clinical teams to focus on decision-making and patient care, rather than
coordination overhead.

large_1blog_thumbnail.webp

medium_1blog_thumbnail.webp

small_1blog_thumbnail.webp

thumbnail_1blog_thumbnail.webp

1blog_thumbnail.webp

How to Prevent Missed Follow-Up on Incidentally Detected Lung Nodules in Hospital Systems

A patient comes in for <a href="https://www.qure.ai/blog/qct-lung-catching-lung-cancer-early">lung cancer
screening</a>. The scan is reviewed, and somewhere in the report, there
it is - a pulmonary nodule.

Sometimes it’s expected. Sometimes it’s
not.

The question is rarely "what is this?" The real questions is "what do i do with this now?"

Most nodules are benign. But a small
number represent early malignancy, and those are the ones that matter. The
challenge is not just identifying nodules, it’s deciding which ones require
action and which ones can be safely followed.

That’s where variability starts to creep
in.

On paper, nodule management seems
straightforward. Guidelines exist. Thresholds are defined. Follow-up intervals
are published.

A 5 mm nodule in one patient may be low risk. The same nodule in another patient may not be.

Growth may be subtle. Measurements may
vary. Prior imaging may not be easily accessible. And sometimes, what is
written in the report doesn’t fully translate into a clear plan.

That’s why managing incidental nodules
during screening is less about memorizing thresholds and more about applying consistent
clinical reasoning.

Start with a framework,
but don’t stop there

Most clinicians rely on Lung-RADS
within screening programs and Fleischner guidelines for incidental
findings outside of them.

These frameworks are essential. They
provide structure and reduce unnecessary variation.

But they are starting points, not
endpoints.

Nodule size is often the first thing we
look at.

Small nodules, particularly those under 6
mm, are usually low risk. Larger nodules require more structured follow-up. But
size alone is not enough.

Two nodules of the same size can behave
very differently.
That’s why management decisions need to be contextual, not formulaic.

If there is one principle that
consistently guides nodule management, it is this:

A single scan provides a snapshot. Serial imaging
provides a pattern.

Stability over time generally suggests benign disease. Growth, even subtle, changes the equation.

This is why follow-up intervals matter so
much. Not just whether follow-up happens, but when it happens.

Too early, and you may not see meaningful change. Too late, and you risk delaying diagnosis.

Subsolid nodules often require a
different approach than solid nodules.

This means follow-up may extend over
years, not months.

Recognizing these differences is
essential. Applying a single approach to all nodules leads to either
over-management or missed risk.

Most nodules will remain under
surveillance.

But certain changes should prompt
escalation:

At that point, the question shifts from “should
we follow this?” to “should we intervene?”

This is where multidisciplinary input becomes important. Decisions around PET imaging, biopsy, or
surgical evaluation are rarely made in isolation.

Even with guidelines, variability remains
a reality.

These are not knowledge gaps. They are execution gaps.

Management decisions are only as
effective as their follow-through.

Even when the right recommendation is
made, it still depends on:

Studies have shown that a significant
proportion of patients with pulmonary nodules do not complete recommended
follow-up (Wiener et al., 2013).

This is where clinical decision-making
meets system design.

At a practical level, managing incidental
nodules well comes down to a few principles:

Each of these steps is straightforward in isolation. The challenge is maintaining consistency across all of them.

As screening programs expand, the number
of nodules requiring follow-up continues to grow.

At that scale, management becomes less
about individual decisions and more about system reliability.

Maintaining visibility across patients,
ensuring follow-up over time, and coordinating across teams requires
infrastructure.

Platforms such as qTrack are designed to
support this layer, helping track patients identified through both <a href="https://www.qure.ai/blog/role-of-an-ai-for-reliable-screening-of-abnormality-in-x-rays">screening and
incidental findings</a> and ensuring that recommended follow-up is
completed. qTrack creates accountability for every nodule and helps ensure that
less to no patients fall through the cracks.

This allows clinical teams to focus on
decision-making, while reducing the risk that patients fall out of the pathway.

Schedule a consultation
with an expert today. Reach out to us on partner@qure.ai.&nbsp;



large_2blog_thumbnail.webp

medium_2blog_thumbnail.webp

small_2blog_thumbnail.webp

thumbnail_2blog_thumbnail.webp

2blog_thumbnail.webp

What Are the Best Practices for Managing Incidental Pulmonary Nodules Detected During Lung Cancer Screening?

The part that breaks
isn’t detection, it’s what happens next

A pulmonary nodule is identified on a
scan. The report is finalized. A follow-up interval is recommended.

And then, in many cases, the patient
disappears from the pathway.

Not immediately. Not obviously. But
somewhere between that initial finding and the next scan, continuity is lost.

This is the gap closed-loop systems are
meant to solve. Not by improving detection, but by ensuring that once a finding
is identified, it continues to move forward until the appropriate next step is
completed.

Not all nodules enter
the system the same way

One of the first things that becomes
clear in practice is that nodules don’t come from a single source.

<a></a>Start by separating two workflows: screening vs incidental

One of the most common operational
mistakes is treating all nodules the same.

In practice, nodules enter the system
through two very different pathways:

These incidental findings are where
most systems struggle. They often lack a defined pathway from the start. And
without that, they are far more likely to be lost.

A reliable follow-up
system doesn’t treat these pathways separately forever. It recognizes the
difference at entry, but brings both into a unified tracking model once
identified.

The mistake of trying to
track everything

It’s tempting to include every<a href="https://www.qure.ai/blog/how-lung-nodule-tracking-software-supports-nurse-navigators"> nodule in a
tracking system</a>. In reality, that approach quickly becomes
unmanageable.

High-functioning programs make an early decision:

what should enter the
system, and what should not.

This usually means defining thresholds
based on:

The goal is not to capture everything. It
is to capture what requires action.

Without this filter,
the system becomes noisy, and meaningful findings are harder to prioritize.

In many setups, follow-up is tied to
static lists, patients who need to be checked at some point.

What works better in practice is a shift
toward <a href="https://www.qure.ai/blog/qure-ai-driving-efficient-workflows">time-driven
workflows</a>.

Every patient should have a clearly
defined “next action date.” Not just a recommendation, but a point in time when
something needs to happen.

As that date approaches, the system
should surface the patient.
&nbsp;If the date passes without action, the
system should escalate.

This changes follow-up from something passive into
something that is actively managed.

Where variability
quietly enters measurement

One of the less discussed challenges in
nodule management is measurement consistency.

A nodule may appear stable in one read
and growing in another, simply due to technique.

Over time, this introduces uncertainty
into decision-making.

Reliable systems don’t just track patients. They also
support:

Another common scenario is the absence of
prior imaging.

A nodule is identified, but there is no
baseline for comparison.
&nbsp;Is it new? Has it been stable for years?
There is no way to tell.

These cases carry a different kind of
risk.

Insights onArtificial Intelligence &Healthcare Technology

Insights on
Artificial Intelligence &
Healthcare Technology