In osteosarcoma the survival gap is made before diagnosis, not after it

Design

Single-institution retrospective comparative study with multivariable Cox regression, level III

Population

88 children, adolescents and young adults with high-grade osteosarcoma, 2010-2024

Primary outcome

Presentation, treatment timelines and survival across five measures of social circumstance

Effect

Median symptom duration 13 vs 6 weeks; five-year overall survival 37% (95% CI 21-53) vs 65% (49-78); deprivation group HR 2.0 (95% CI 1.1-3.9); no difference in time from diagnosis to treatment

Eighty-eight children, adolescents and young adults with high-grade osteosarcoma diagnosed at a single institution between 2010 and 2024 were classified by five different measures of social circumstance: Area Deprivation Index, Childhood Opportunity Index, insurance type, census tract poverty level, and self-reported race or ethnicity.

Only the deprivation index separated anything. Patients in the high-deprivation group presented through urgent or emergency care in 51% of cases against 22% (OR 1.9, 95% CI 1.2-2.9; P = 0.006), and had carried symptoms for a median 13 weeks against 6 (P = 0.002). Their five-year overall survival was 37% (95% CI 21-53) against 65% (95% CI 49-78; P = 0.04), and five-year local recurrence-free survival 81% against 96%. On multivariable Cox regression, deprivation group (HR 2.0, 95% CI 1.1-3.9) and metastatic disease at presentation (HR 3.2, 95% CI 1.7-6.4) were independently associated with death. None of the other four measures identified any survival difference.

And the time from diagnosis to starting and completing treatment did not differ by any measure. That is the finding that should change what clinicians do. Once these patients reached the sarcoma service they were treated identically; the disadvantage was entirely accumulated in the seven extra weeks before anyone made the diagnosis. The lever is not in the oncology pathway. It is in what happens when an adolescent presents to a general clinic or an emergency department with persistent bone pain - and in a system like India's, where first contact is even more dispersed and imaging less immediate, that seven-week gap is the part worth attacking.

• Image persistent unexplained bone pain in an adolescent at first presentation rather than at review
• Treat night pain, rest pain or a palpable mass as a reason to radiograph today, not to observe
• Record symptom duration explicitly at referral - it was the variable that tracked survival
• Do not read a reassuring examination as a reason to defer a plain film; the gap here was pre-imaging
• Give general practitioners and emergency staff a direct route to imaging, not only to a clinic appointment

Why it matters

Treatment was identical once these patients arrived, so every part of the disparity was created before the diagnosis was made.

Don't overread it

Retrospective and single-institution: it shows delay and worse survival in the same group, not that shortening the delay would fix it.

The statistics, in plain English

The striking part of this analysis is the disagreement between measures. Four indices of social circumstance found nothing and one found a doubling of mortality risk, in the same 88 patients. That is not because deprivation does not matter but because these instruments measure different things - the deprivation index weighs income heavily, the others do not - and a study that picks the wrong one reports a null. With 88 patients the confidence intervals are correspondingly wide: the hazard ratio of 2.0 runs from 1.1 to 3.9, so the direction is established and the magnitude is not. Single-institution and retrospective, this cannot show that faster diagnosis would have changed survival, only that the delay and the deaths occurred in the same group.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.