Oncology imaging carries more operational weight than almost any other workflow a PACS has to support. A single study is not one acquisition but several, layered together into one exam. The read that follows is rarely a single study either: an oncology radiologist is routinely working across a current PET-CT, a prior PET-CT, a diagnostic CT, and sometimes an MRI, in the same sitting, because comparison across time and modality is the actual clinical task.
And because cancer patients get imaged repeatedly for years, none of that imaging history can be treated as disposable. What looks like a storage problem at first glance is really a retrieval and retention problem, and it shows up first in how long a radiologist waits before a read can even start.
This guide covers what makes PET-CT and multi-modal oncology reading structurally different from a routine radiology workflow, and what that difference demands from the PACS underneath it. It is one of several specialty workflows where imaging infrastructure choices matter for reasons specific to the specialty rather than to volume alone, alongside cardiology’s cath lab and echo demands and mammography’s tomosynthesis storage curve.
Why a PET-CT Study Is Not One Series
A routine CT or MRI study is, in DICOM terms, a fairly contained object: one acquisition, a handful of series, one exam a PACS stores and retrieves as a single unit. A PET-CT study does not work that way. It starts as two separate acquisitions, the CT and the PET, captured on the same gantry in the same visit, and it does not stay two.
The CT does double duty: it produces its own diagnostic series, and it also supplies the attenuation map the PET reconstruction needs to correct for tissue density. That correction step is where a PET-CT study picks up series a routine exam never has.
Attenuation-Corrected and Non-Corrected Series
PET data collected straight off the detector is distorted by how deep in the body a lesion sits and how much tissue the signal has to pass through before it is captured. Attenuation correction fixes that using the paired CT, and PACS systems typically keep the corrected and uncorrected PET reconstructions as separate stored series rather than discarding one, since a radiologist or physicist may need to check the non-corrected images against artifacts the correction step can introduce. That alone puts three series in a single PET-CT study before counting the CT itself.
Multiple Reconstructions, One Study
Reconstruction rarely stops there. A department may generate additional series for specific reading needs, a fusion overlay series, a different slice thickness, sometimes a second reconstruction pass for comparison, and each one is stored inside the same study rather than as a separate exam. None of that shows up on the worklist as extra cases.
It shows up as one study that is structurally heavier than a routine CT or MRI, which is why PET-CT sizing does not follow the same modality math as the rest of a department’s archive. For the underlying storage and bandwidth arithmetic that a heavier study like this feeds into, see PACS capacity planning for storage, bandwidth, and compute, which walks through how per-modality study size turns into an annual footprint.
What Multi-Modal Reading Actually Requires
An oncology read is rarely a single study opened cold. It is usually the current PET-CT set against the prior PET-CT to see what has changed, plus a diagnostic CT from the same encounter or a recent MRI when the anatomy calls for it.
Comparison across time and modality is not a supplementary step in that read. It is the read. A radiologist forming a staging or restaging impression is doing it by holding several studies against each other, not by reading one in isolation.
Comparing Across Time, Not Just Across Series
Oncology response frameworks formalize what this workflow has always required operationally: a current study assessed against a documented baseline and, at each follow-up, against the most recent prior. The RECIST criteria, the standard most widely used for tracking lesion-size change on CT or MRI, and a parallel metabolic framework called PERCIST for PET, both depend on the same underlying requirement: the comparison study has to actually be available and correctly matched to the current one when the read happens.
Neither framework is a workflow instruction. Both are a reminder that the infrastructure question comes first. If the prior is not there, or takes several minutes to surface, the comparison cannot happen the way the read depends on it happening.
Registration and Alignment Between Studies
Comparing two PET-CT studies, or a PET-CT against a diagnostic CT, only works cleanly if the studies actually line up. Patient position shifts between visits, organs move, and a lesion that sits in one location on a CT slice needs to correspond to the same lesion on the PET overlay and on the prior study a radiologist is checking it against. Registration and alignment between studies is what makes that correspondence usable rather than approximate, and it is a workflow expectation the PACS has to support directly rather than leave to the radiologist to reconcile by eye. A 2024 multidisciplinary expert-consensus paper, covering PET-CT imaging workflow for long-axial-field scanners, documents imaging reconstruction and review protocols for these scanners in detail alongside the acquisition workflow itself, a reminder that the reading tools around a PET-CT study carry as much weight as the acquisition.
What This Demands From a PACS
None of the above works if the PACS makes a radiologist go looking for it. Fast prior retrieval, series-level navigation inside a single heavy study, and registration between current and prior exams need to happen inside the same worklist context, not through a manual search across a separate archive. OmniPACS builds worklists around the patient rather than the originating study, so a current PET-CT lands next to its prior PET-CT and any diagnostic CT or MRI from the same care episode in one view, instead of requiring a radiologist to open and cross-reference three separate lookups mid-read.
Longitudinal Follow-Up Is a Retention and Retrieval Problem
Cancer patients are not imaged once. Staging, treatment-response checks, and post-treatment surveillance can mean PET-CT and cross-sectional imaging repeated every few months for years, sometimes a decade or more past initial diagnosis. Every one of those studies is a potential prior for the next read, which means none of them can be treated as safe to archive out of easy reach the way an unremarkable one-off exam might be.
Retention Windows Built for Oncology’s Timeline
That reality changes what a retention policy needs to account for. A department can build tiered storage that moves older, less-accessed studies to cheaper archive tiers, the way multi-site PACS architecture generally handles hot, warm, and cold data, but an oncology-heavy archive cannot assume old automatically means unlikely to be needed. A five-year-old PET-CT on a patient now in year six of surveillance is exactly the kind of prior a radiologist needs fast, not the kind that is safe to push to slow, minutes-to-restore cold storage. OmniPACS is built to scale storage capacity with actual usage rather than a fixed ceiling set years in advance, which matters for an oncology-heavy service line specifically because volume and retention pressure both keep growing for years after a patient’s original diagnosis.
What Slow or Missing Priors Cost a Department
The cost of getting this wrong is not abstract. A 2026 analysis of Medicare claims data found a 27 percent increase in the time between scan and interpretation between 2023 and 2024 alone, part of a broader trend of imaging interpretation turnaround times climbing across radiology. That pressure lands hardest on studies with the most built-in comparison work, and an oncology read against multiple priors and multiple modalities is exactly that kind of study.
A prior that is slow to load, or one a radiologist has to hunt for across a separate legacy archive, adds friction to a read that already carries more comparison work than a routine exam. Multiply that across a full oncology worklist and the delay stops being one radiologist’s bad afternoon. It becomes a department-wide drag on staging and restaging turnaround, one that can directly affect how quickly a treatment decision gets made downstream.
What to Look For in a PACS for Oncology Volume
Departments with meaningful oncology or PET-CT volume should evaluate a PACS against what this workflow actually requires, not the general feature list:
- Fast retrieval of prior PET-CT and cross-sectional studies, not a multi-minute pull from a separate archive
- Series-level navigation that surfaces attenuation-corrected, non-corrected, and fusion series without manual sorting
- Registration and alignment tools that let a radiologist compare current and prior studies without eyeballing the correspondence
- A worklist built around the patient across modalities, rather than one that requires separate logins per acquisition source
- Retention and tiering policies that account for oncology’s multi-year follow-up pattern rather than a generic age-based cutoff
- Confirmed throughput for PET-CT’s heavier, multi-series studies, tested against real volume rather than average study size
Facilities weighing where a current PACS falls short on any of the above can explore OmniPACS Solutions to see how a cloud PACS handles prior retrieval and retention specifically for oncology-heavy imaging volume.
Where a General-Purpose Cloud PACS Fits
It is worth being direct about this: OmniPACS is a general-purpose cloud PACS, not an oncology-specific platform. It does not replace treatment-planning software, does not run tumor board conferencing, and does not calculate RECIST or PERCIST scores itself. What it provides is the imaging infrastructure underneath those tools: DICOM storage and retrieval sized for PET-CT’s multi-series studies, fast prior access across a years-long follow-up history, and a worklist that keeps a patient’s full imaging record, across PET-CT, CT, and MRI, in one place instead of scattered across separate systems.
For an oncology service line already running RECIST- or PERCIST-based response assessment and its own tumor board process, that infrastructure layer is the piece worth confirming before volume grows past what the current archive was sized for. Getting the imaging layer right does not replace the clinical and oncology-specific tools built on top of it, but it determines whether those tools have fast, complete data to work with in the first place.
Frequently Asked Questions
What is attenuation correction in a PET-CT study?
Attenuation correction adjusts raw PET data for signal loss caused by tissue depth and density, using the paired CT to make that correction. PACS systems typically store the corrected and uncorrected PET reconstructions as separate series in the same study, part of why a PET-CT study holds more series than a routine CT or MRI.
Can a general-purpose PACS handle oncology and PET-CT imaging?
It can handle the imaging infrastructure: DICOM storage for fused, multi-series PET-CT studies, fast prior retrieval, and a worklist that groups a current PET-CT against prior studies and other modalities. It does not replace oncology-specific software like treatment-planning or tumor-board platforms, which sit downstream of the imaging layer.
What does multi-modal reading mean in oncology imaging?
Multi-modal reading means comparing a current study against more than one prior source in the same session, typically a prior PET-CT plus a diagnostic CT and sometimes an MRI. That comparison is the read itself, which is why fast retrieval and series-level navigation across studies matter more in oncology than in most other imaging workflows.
How much storage does a PET-CT study need?
More than a routine CT or MRI, though the exact figure depends on scanner and reconstruction settings. A fused PET-CT study is not one series but several: the PET acquisition, the paired CT, and at least two PET reconstructions, corrected and uncorrected for attenuation, a structural reason the study runs larger.
Oncology imaging is not a heavier version of general radiology so much as a different kind of read altogether, one built on comparison across time and modality rather than a single study in isolation. Getting the PACS layer right, so a multi-series PET-CT and years of priors stay fast to retrieve, is worth the evaluation time for any department where oncology makes up a meaningful share of volume. Departments ready to compare deployment models can review flexible pricing for every need to see how storage costs track with oncology imaging volume rather than a fixed footprint sized for last year’s caseload.