Trastuzumab deruxtecan leads the network, with a fivefold pneumonitis signal attached

Design

Bayesian network meta-analysis of phase 2 and 3 randomised trials, PRISMA-NMA compliant

Population

3,339 patients with HER2-positive metastatic breast cancer after trastuzumab-based therapy

Primary outcome

progression-free and overall survival, confirmed response and safety

Effect

trastuzumab emtansine vs deruxtecan: PFS HR 3.30 (95% CrI 2.06-5.28), OS 1.52 (1.19-2.01); pneumonitis OR 5.55 (2.79-12.16) for deruxtecan

Seven independent randomised trial families and 3,339 patients with HER2-positive metastatic breast cancer beyond trastuzumab-based therapy were combined in a Bayesian network meta-analysis of second-line and later regimens.

Against trastuzumab deruxtecan, trastuzumab emtansine performed worse on progression-free survival (hazard ratio 3.30, 95% credible interval 2.06-5.28) and overall survival (1.52, 1.19-2.01). Confirmed response favoured deruxtecan over lapatinib-capecitabine (odds ratio 7.43), pyrotinib-capecitabine (3.35) and emtansine (6.31). But the safety networks do not follow: any-grade adverse events showed no advantage, serious adverse event comparisons were inconclusive, and the odds of interstitial lung disease or pneumonitis were 5.55 times higher than with emtansine (2.79-12.16).

Two structural limitations are load-bearing rather than decorative. Deruxtecan was not represented in the second-line-only overall survival network, so its survival ranking comes from mixed-line evidence; and it was absent from the Asia-dominant networks, which is exactly the evidence base an Indian oncologist would want. Indirect comparison assumes the trials are similar enough to borrow strength across, and across seven trial families spanning different lines and regions that assumption is doing real work. Treat this as confirming a hierarchy already visible in the individual trials, with the pneumonitis risk as the number to carry into the consultation.

• Counsel explicitly about interstitial lung disease before starting trastuzumab deruxtecan.
• Investigate new or worsening cough or breathlessness on deruxtecan as pneumonitis until excluded.
• The survival ranking draws on mixed-line trials, not second-line-only evidence.
• Asia-dominant trials did not contribute to the survival networks here.
• No safety advantage over comparators was demonstrated on any-grade or serious adverse events.

Why it matters

It sets the ranking most second-line HER2 decisions already follow, and attaches the toxicity number that should be said out loud when they are made.

Don't overread it

Indirect comparison across trial families, with the leading drug absent from two of the survival networks.

The statistics, in plain English

Network meta-analysis compares treatments that were never tested against each other by routing through common comparators, which requires the trials to be similar enough for that route to be valid. Here they span different lines of therapy and different regions, and the authors say plainly where deruxtecan was missing from a network - a caveat that means those particular rankings rest on indirect evidence alone. The pneumonitis credible interval, 2.79 to 12.16, is wide but nowhere near 1.0: the risk is real, its size is not precisely known.

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.