Topic 03 · Equipment & capital · Deep dive

The hardware divide: who owns the machines, how old they are, what they cost

High-income countries field roughly forty CT scanners per million people; low- and middle-income countries field fewer than one. In nuclear imaging the gap widens to three orders of magnitude. Meanwhile the rich world's own fleets are ageing, and the top of the market has entered its biggest technology transition in a generation.

40×
CT density gap: ~40 scanners per million in high-income countries vs <1 in LMICs
IAEA IMAGINE / EJRO
457×
PET density gap: 3.2 per million (high-income) vs 0.007 (low-income). Nine of Africa's 54 countries had any PET scanner; by 2026 that had risen to 12
IAEA NUMDAB / Clin Transl Imaging, 2020
33%
of Canada's CT fleet is over 10 years old (average age 8.2 years; run 14 h/day)
CADTH CMII 2022–23
2M+
patients scanned on photon-counting CT since its 2021 debut — the segment the high end is converting to
Siemens Healthineers, 2025
Global fleet density

Scanners per million people: a map of capital, not need

Density tracks wealth and payment systems far more than disease burden. Japan owns the densest CT fleet on earth and, notably, among the fewest radiologists per person to read its output.

CT scanners per million

Latest available (2022–2024).
Japan2024 · densest fleet on earth
120
Australia2024
74
High-income average~40 per million
40
OECD average2022–23
30
Canada2022–23 · 30th of 34 OECD
14
Low/middle-income countriesIAEA IMAGINE
1
Sources: OECD via Statista (2024); CADTH CMII 2022–23; IAEA IMAGINE via Eur J Radiol Open (HIC ~40/M, LMIC <1/M). The US reports 86 CT+MRI+PET units per million combined vs an OECD average of 51 (OECD Health at a Glance 2025).

MRI units per million

Latest available (2020–2022).
United States2020
35
Germany2022 · age-adjusted
32
Canada2022 · 27th of 31 universal systems
11
Israel2020 · works each unit hardest
5.5
Sub-Saharan Africa (typical)CAMERA / WHO · many countries ≤1
0.8
Sources: OECD 2011–2020 series (Martella et al.); Fraser Institute (age-adjusted); CAMERA MRI-in-Africa needs assessment / WHO. Across West Africa there is as little as one MRI scanner per country. Japan and Greece report the world's highest MRI densities.
Nuclear imaging cliff
SPECT: 17.9 vs 0.036 cameras per million (high- vs low-income) — a ~500× gap across the world's 25,000+ cameras, with 134 of 195 countries having any nuclear-medicine facility (IAEA, 2020). IAEA's IMAGINE database now counts ~27,000 systems across >140 countries.
Catch-up arithmetic
To reach high-income equipment levels, LMICs would need to add +11.4 CT and +5.2 MRI units per million people — thousands of installations, before staffing any of them.
The human cost
Scaling imaging up across LMICs would avert an estimated 2.46 million cancer deaths over 2020–2030 — 3.2% of cancer deaths, and 54.9M life-years — at a cost of $6.84bn. A modelled projection, and the Lancet Oncology Commission's headline case for closing the hardware gap.
Nuclear imaging

PET and SPECT: the steepest gradient in medicine

Nuclear imaging density by income level (log-scale territory)

Cameras/scanners per million population. The low-income bars are real values, not rendering errors.
SPECT · high-income17.9 per million
18
SPECT · low-income0.036 per million
0.0
PET · high-income3.2 per million
3.2
PET · low-income0.007 per million
0.0
Source: IAEA NUMDAB/IMAGINE via Paez et al., Clin Transl Imaging 2020. One growth story: Latin America went from 22 PET systems (2005) to 301 (2019) — a 21% CAGR — and still sits at only 0.47 per million.
What the machines cost

From $55k as-is to $5M photon-counting

The price ladder now spans two orders of magnitude within CT alone — and the refurbished market, at roughly half of new pricing, is how much of the world buys in.

SystemNew (installed)RefurbishedNotes
CT — entry (16–64 slice)$350k–$750k$80k–$250kRefurb typically ~50% below new; "as-is" units from ~$55k
CT — premium single-source$1M–$2M$250k–$450kHigh slice counts, cardiac packages
CT — dual-source photon-counting$2M–$5M+No secondary market yet; top of the range
MRI — 1.5T (workhorse)from ~$500k
(projects €400–550k)
$150k–$500k70 cm bore ≈ +25% vs 60 cm; installation $50k–$80k
MRI — 3T$1M–$3M$400k+Roughly double 1.5T pricing tier
Ultrasound≈ one-tenth of CT/MRIfrom low $10ksThe access technology: cheap to buy, simple to maintain

Indicative US/EU market ranges from dealer and vendor guides (2023–2026); not audited transaction data. Service contracts, shielding, and site works add materially: MRI annual service alone commonly runs into six figures over a system's life.

The helium line-item
A conventional superconducting magnet holds roughly 1,000–2,000 L of liquid helium (~1,500 L typical). Refill pricing is volatile and contract-dependent — trade sources cite ~$32–35/L, with sharp spikes during the 2022 shortage. Philips puts the cost of a single quench at up to €100,000.
The helium exit
Philips has installed 2,000+ sealed BlueSeal 1.5T magnets since 2018 — saving a claimed 6 million litres of helium. Low-field 0.55T "High-V" systems attack the same dependency from below.
The detector revolution
Photon-counting CT — called by the FDA the biggest CT advance in a decade — has scanned 2M+ patients since 2021. Siemens is exiting conventional CT at the high end; GE, Canon and NeuSoft are entering the segment.
Fleet age & utilisation

The rich world's quiet problem: old machines, worked hard

Canada's CT fleet by age

Share of installed units, 2022–23. Average age: 8.2 years.
0–5 years old29.8% of units
30%
6–10 years old36.7% of units
37%
Over 10 years old33.4% of units
33%
Source: CADTH Canadian Medical Imaging Inventory. Units operate an average of 14 hours per day. Older platforms typically lack the newest dose-reduction and AI-reconstruction capabilities — fleet age is a quality issue, not just a capital one.

The basic-imaging benchmark

General radiography units per million vs the WHO's recommended 20.
WHO recommendationbasic access threshold
20
Tanzania · private sector2016 registry
25
Tanzania · public sectorserves most of the population
5.7
Source: Tanzanian Atomic Energy Commission registry analysis (PMC). Tanzania's public sector runs one CT per ~70 radiography units — against a suggested 1:10 — a planning dilemma common to the 33 sub-Saharan countries with similar health budgets.

Machines follow money, not disease

Equipment density is the cleanest proxy for imaging capital, and it maps almost perfectly onto national income: roughly 40 CT scanners per million people in high-income countries against fewer than one across the low- and middle-income world, with the gradient steepening through MRI (a 3.6× spread even inside the rich world) to nuclear imaging, where the high-to-low-income ratio reaches three orders of magnitude. Nine African countries in fifty-four have a PET scanner of any kind. The Lancet Oncology Commission's arithmetic gives the gap its human denominator: about 2.46 million cancer deaths a year attributable to absent imaging.

Density is not capacity

Japan is the standing rebuttal to hardware-first thinking: the world's densest CT and MRI fleets alongside one of the OECD's thinnest radiologist workforces. Israel is the counter-example in the other direction — among the fewest MRI units per person in the OECD, worked at ~8,400 exams per unit per year, six times Greece's rate. The scanner count tells you what a system bought; the exams-per-scanner and reports-per-radiologist figures tell you what it actually delivers. Any equipment investment case that ignores the workforce page of this site is half a business case.

The refurbished market is global access infrastructure

With new CT at $350k–$2M and MRI projects starting around half a million dollars, the certified-refurbished channel — reliably about 50% below new — is not a niche: it is how large parts of the world acquire cross-sectional imaging at all. Ultrasound sits below the whole ladder at roughly a tenth of CT/MRI cost with trivial maintenance, which is why it anchors every serious LMIC access strategy. The WHO's 20-radiography-units-per-million benchmark is a reminder that for much of the world the capital question is still basic X-ray, not photon counting.

A generation change at the top

The top of the market is mid-transition on two axes at once. Detectors: photon-counting CT has gone from first install (2021) to 2M+ patients and a multi-vendor race, with the incumbent abandoning conventional technology in its high-end segment entirely. Cryogens: 2,000+ sealed helium-free magnets are already in the field, a direct response to the helium fragility documented in Topic 07. Both transitions will eventually cascade down the price ladder and into the refurbished market — the mechanism by which today's $5M flagship becomes the 2035 workhorse.

On the data. Scanner-density definitions differ across sources (installed vs operational; public-only in Canada's inventory; age-adjusted in the Fraser compilation) and years span 2016–2024. IAEA IMAGINE aggregates voluntary submissions of uneven quality. The nuclear-medicine density figures (SPECT 17.9 vs 0.036/M; PET 3.2 vs 0.007/M; 134 countries; 9 of 54 African countries with PET) come from Paez et al. via NUMDAB and describe the position around 2020; IAEA's IMAGINE database now reports roughly 27,000 SPECT systems across more than 140 countries, and 12 African countries offering PET/CT — so these ratios understate recent growth, though not the scale of the gap. All prices are indicative dealer/vendor ranges, not audited transactions, and vary widely with configuration and service terms. Helium refill pricing in particular is contract-dependent and swung sharply during the 2022 shortage; the quoted range is a trade estimate, not a spot price. Installed-base figures for photon-counting CT and BlueSeal are vendor-reported. The 2.46 million figure is a modelled projection of deaths averted by an imaging scale-up over 2020–2030 (Lancet Oncology Commission) — not a count of deaths currently attributable to lack of access.

Sources

  1. OECD CT scanner density by country (via Statista, 2024)
  2. CADTH Canadian Medical Imaging Inventory 2022–23 — units, fleet age, utilisation
  3. EJRO — improving imaging access in LMICs (IAEA IMAGINE gap analysis: 40:1 CT, +11.4 CT/M, +5.2 MRI/M)
  4. Paez et al., Clin Transl Imaging 2020 — global nuclear medicine (PET/SPECT density by income)
  5. Clinical Radiology 2024 — imaging inequality (West Africa MRI)
  6. Lancet Oncology Commission 2021 — imaging scale-up would avert 2.46M cancer deaths (2020–2030); $6.84bn cost, 54.9M life-years
  7. CAMERA — MRI in Africa needs assessment
  8. Tanzania equipment registry analysis — the diagnostic divide
  9. CT price guide (Cassling)
  10. MRI price guides & helium economics (Block Imaging / DirectMed: ~$32–35/L)
  11. Philips BlueSeal — ~2,000 installations, >6M litres of helium saved; a quench can cost up to €100,000
  12. Science & Medicine Group 2026 — photon-counting market, BlueSeal installed base
  13. Applied Radiology 2025 — PCCT 2M+ patients scanned