For most imaging centers and small hospital departments, the math in 2026 is straightforward: a full-time in-house radiologist is a fixed cost that recent physician compensation surveys put well above $500,000 per year fully loaded, while outsourced teleradiology is a variable cost billed per report, typically tens of dollars for an X-ray up to low hundreds for advanced cross-sectional studies at industry-typical rates. Below roughly 40 to 60 billable studies per day of steady volume, per-report outsourcing is almost always cheaper; above that, an in-house reader starts to win on unit cost but loses on coverage, subspecialty breadth, and volume spikes. That is why the model most facilities actually run is neither pure in-house nor pure outsourced: it is a hybrid, with an in-house core for daytime volume and a teleradiology partner like AstraRad handling nights, weekends, overflow, and subspecialty gaps at published SLA tiers of STAT under 1 hour, Urgent under 4 hours, and Routine under 24 hours.
This guide walks through the real numbers on both sides, the trade-offs that vendors usually gloss over, and a break-even framework you can apply to your own volume.
What does an in-house radiologist actually cost in 2026?
The salary line is only the start. A realistic fully loaded cost model for one in-house radiologist FTE includes:
| Cost component | What it covers | Typical share of total |
|---|---|---|
| Base compensation | Industry surveys consistently place radiologist pay among the highest of all specialties, commonly cited above $500,000 | 65 to 75 percent |
| Benefits and payroll taxes | Health insurance, retirement match, employer taxes, malpractice coverage | 15 to 25 percent of base |
| Recruiting | Search firm fees, signing bonuses, relocation, in a market with a persistent radiologist shortage | Often a six-figure one-time cost, amortized |
| Call and coverage premiums | Night and weekend call pay, or the cost of a second hire to make call sustainable | Highly variable |
| Non-productive time | Vacation, CME, sick leave, administrative duties: weeks per year when the salary runs but no reports are signed | 10 to 15 percent of paid time |
Two structural problems compound the raw cost. First, one radiologist cannot cover 168 hours a week; realistic solo coverage is closer to 45 to 50 reading hours, which means nights, weekends, and vacations still need a coverage answer. Second, one radiologist is one set of fellowship training. A general radiologist reading everything from pediatric ultrasound to cardiac MRI is the norm in small departments, and it is exactly the pattern the subspecialty literature identifies as the highest-miss configuration.
None of this means in-house staffing is a mistake. It means the true comparison is not "salary vs per-report fee." It is "fully loaded fixed cost plus a separate coverage solution" vs "variable cost with coverage included."
What does outsourcing actually cost?
Teleradiology pricing across the industry falls into three broad structures: per-report fees, monthly minimums or retainers, and platform or technology fees layered on top. Industry-typical per-report rates run from roughly $10 to $25 for plain film X-ray, $30 to $80 for CT depending on complexity, $40 to $100 for MRI, and more for PET-CT and cardiac studies, with STAT priority usually carrying a multiplier. Those are market ranges, not AstraRad's prices; for a full breakdown of how the industry prices reads, see our guide to teleradiology costs.
AstraRad's structure is deliberately simple: per-report pricing, no minimums, no subscriptions, no platform fees, and the priority multiplier printed on the rate card. Send one study a month or ten thousand. Rates vary by study type across the full menu we read, from CR X-ray and ultrasound through mammography, single and multi-region CT, coronary CTA, MRI including cardiac, nuclear medicine gamma studies, and PET-CT. Request a rate card through our contact page and you will have it within one business day; the pricing page explains the model.
The economic property that matters for this comparison is the shape of the cost curve, not any single price. Per-report outsourcing costs zero when volume is zero and scales linearly with volume. An in-house FTE costs the same in a slow February as in a record October. Which curve wins depends entirely on your volume, which is what the break-even section below quantifies.
The honest trade-offs: where each model wins
Ranking pages on this topic are mostly vendor sales copy. The academic literature on teleradiology outsourcing is more critical, and some of its criticisms are fair. Here is the balanced version.
Where in-house wins
- Hallway access. A clinician can walk into the reading room, pull up a study, and talk through it. No teleradiology workflow fully replaces that, even with direct phone lines to the reading radiologist.
- Procedures and on-site presence. Fluoroscopy, image-guided biopsies, contrast reaction coverage, and technologist supervision require a physically present radiologist. Teleradiology reads images; it does not do procedures.
- Institutional knowledge. An in-house radiologist learns your referrers, your scanner quirks, your patient population, and your prior-study history in a way a distributed panel builds more slowly.
- Unit cost at very high steady volume. Past the break-even point, each additional in-house read is nearly free, while each outsourced read still carries its fee.
Where outsourcing wins
- Coverage. A per-report partner covers 24/7/365. AstraRad reads across 12 time zones, so a 2 a.m. head CT in Ohio is daytime work for a rested radiologist somewhere on the panel, not the last case of an exhausted overnight shift.
- Subspecialty breadth. No small department can employ neuro, MSK, body, cardiac, breast, pediatric, chest, ER, oncologic, and nuclear imaging subspecialists. AstraRad's panel of 240 board-certified, fellowship-trained subspecialists routes every study to a reader trained in that anatomy.
- Elastic capacity. Volume spikes, scanner additions, and backlogs do not require a hire. AstraRad carries headroom for 25,000 additional studies per month, and a backlog of 8,000 studies clears in under 30 days. See how facilities use this for overflow radiology reads.
- Measured, contractual turnaround. In-house turnaround is whatever the day allows. Outsourced turnaround is a contract: AstraRad's tiers are STAT under 1 hour, Urgent under 4 hours, Routine under 24 hours, with a 28-minute median STAT turnaround and 99.4 percent SLA compliance over the trailing 12 months, published on our SLA page.
The hidden costs the literature documents, and how to avoid them
Critical academic coverage of radiology outsourcing flags real failure modes, and they are worth taking seriously:
| Hidden cost | Where it comes from | How to avoid it |
|---|---|---|
| Double reading expense | Preliminary-only nighthawk reads that a local radiologist must overread and re-sign in the morning | Contract for final signed reports at every tier. AstraRad delivers final reads, including overnight |
| Communication friction | Critical findings routed through call centers instead of direct clinician contact | Require a documented critical-findings communication policy in the SLA |
| Quality opacity | Vendors that publish no discrepancy or QA data | Ask for measured numbers. AstraRad independently double-reads 1 in 20 reports and holds a major discrepancy rate under 0.3 percent |
| Integration projects | Platform fees and months-long IT engagements before the first read | Choose DICOM-native routing. AstraRad accepts studies by DICOM from your PACS or portal upload, with no integration project, and delivers reports via HL7 or FHIR |
| Licensing gaps | Readers not licensed where your patients are | Verify that the signing radiologist is licensed in the state where your patients are located, on every report |
If a vendor cannot answer these five points in writing, that tells you something. Our guide on how to choose a teleradiology company turns this into a full evaluation checklist.
The hybrid model: what most facilities actually run
The in-house vs outsourced framing is a false binary. The dominant real-world model is a hybrid:
- In-house core reads weekday daytime volume, performs procedures, supervises contrast, and owns clinician relationships.
- Teleradiology partner covers nights, weekends, and holidays, absorbs overflow when the in-house worklist backs up, and provides subspecialty reads the core group lacks, such as cardiac MRI, coronary CTA, or PET-CT.
The hybrid works because it assigns each model the work it is structurally best at. The in-house team stops taking overnight call, which is the single biggest driver of radiologist burnout and turnover, and the facility stops paying fixed cost for coverage hours that carry a fraction of daytime volume. Per-report pricing with no minimums is what makes the hybrid economical: weekend and holiday coverage costs exactly what the weekend volume costs, and a quiet weekend costs almost nothing.
Hybrid arrangements also de-risk staffing transitions. When an in-house radiologist retires or leaves, the teleradiology partner absorbs the volume during the 12-plus months a replacement search typically takes, without a locum tenens premium.
Break-even framework: when does volume justify hiring?
Here is a framework you can run with your own numbers. All figures below are illustrative industry-typical assumptions; substitute your actual costs and your actual rate card.
- Compute the fully loaded annual cost of the hire. Base compensation plus 20 to 30 percent for benefits, taxes, and malpractice, plus amortized recruiting, plus the cost of covering that reader's nights, weekends, and time off. For most US markets in 2026 this lands between $600,000 and $750,000.
- Compute realistic annual read capacity. A full-time reader signing across a mixed worklist typically produces on the order of 12,000 to 18,000 reads per year after vacation, CME, and non-reading duties.
- Compute the blended per-report cost of outsourcing the same mix. Weight your modality mix against the vendor's rate card. A plain-film-heavy imaging center blends far lower than a CT and MRI-heavy hospital list.
- Divide. Fully loaded cost divided by blended per-report rate gives the annual volume at which the hire breaks even. At a $650,000 loaded cost and a $45 blended rate, that is roughly 14,400 reads per year, about 55 reads per working day, sustained, in the modality mix the hire can actually cover.
Three corrections push the practical break-even higher than the raw division suggests. The hire only breaks even on volume within their own competence, so subspecialty studies still go out. The hire does not solve nights and weekends, so coverage cost remains. And volume must be steady: a facility that hits 55 reads per day in winter and 35 in summer pays for the gap all year.
The honest conclusion: if you have high, steady, general-radiology daytime volume and need on-site procedures, hire, and hire happily. Route everything else, nights, weekends, spikes, and subspecialty work, to a per-report partner. That is the configuration AstraRad is built for, from single-study senders to groups moving thousands of studies a month, drawn from a panel that signs 600,000 reads a year.
Run your own numbers
The fastest way to test the framework is with a real rate card against your last 12 months of volume by modality. Request AstraRad's rate card through our contact page; it arrives within one business day, priced per report by study type with the priority multiplier printed on it, no minimums and no platform fees. Compare the blended cost against your loaded staffing cost, and the right mix for your facility usually becomes obvious in one spreadsheet.