Immunotherapy and the liver: grade it, read the pattern, know when to add a second drug
Woman in her 40s on maintenance pembrolizumab for cervical cancer. AST 235, ALT 144. On methylprednisolone the transaminases fall every day, but the alkaline phosphatase climbs from 126 to 238 and the ammonia from 37 to 93. Is the steroid working?
Checkpoint inhibitor: PD-1, PD-L1 or CTLA-4 antibody (pembrolizumab, nivolumab, ipilimumab). Grade: how many times above the upper limit of normal (ULN); see the table below. Pattern: hepatocellular when ALT leads, cholestatic when ALP leads, mixed in between; the R ratio decides. Steroid-refractory: no meaningful improvement after about 3 days of full-dose steroid. MMF: mycophenolate mofetil.
How to grade (CTCAE version 5, used by ASCO and NCCN)
| Grade | ALT or AST | Total bilirubin | ALP | Checkpoint inhibitor | Steroid |
|---|---|---|---|---|---|
| 1 | Up to 3x ULN | Up to 1.5x | Up to 2.5x | Continue, labs weekly | None |
| 2 | Over 3 to 5x | Over 1.5 to 3x | Over 2.5 to 5x | Hold until grade 1 or less and prednisone 10 mg or less | Prednisone 0.5 to 1 mg/kg if not better in 2 to 3 days |
| 3 | Over 5 to 20x | Over 3 to 10x | Over 5 to 20x | Stop; ASCO says permanently | Methylprednisolone 1 mg/kg IV |
| 4 | Over 20x | Over 10x | Over 20x | Stop for good, admit | Methylprednisolone 1 mg/kg IV, now |
Grade by the worst value of ALT, AST and bilirubin. Abnormal baseline (liver metastases): CTCAE grades by multiples of the patient's own baseline instead of ULN.
Algorithm
What the studies show
| Study | Who | Finding | Number to remember |
|---|---|---|---|
| Li, Hepatology 2022 | 215 patients with ALT over 200, Boston | 1 mg/kg methylprednisolone did as well as 1.5 to 2 mg/kg for the liver, with fewer complications | 18% vs 7%infections, high vs low dose |
| Holmstroem, J Immunother Cancer 2026 | 34 patients, grade 3 to 4, prospective, all biopsied | 59% responded to steroids; 18% needed MMF; 23% relapsed on taper. Cholestatic pattern had the worst outcome | 23%steroid-dependent |
| Alouani, Eur J Cancer 2023 | 130 grade 3 or higher, Gustave Roussy | 8.5% steroid-refractory; MMF brought 9 of 11 back to grade 1 or better | 9 of 11respond to MMF |
| Romanski, Eur J Cancer 2020 | 521 melanoma patients, Denmark | One in three would be graded differently by ALT or by AST; 35.5% relapsed during taper | 35%graded differently by one enzyme |
Bottom line: 1 mg/kg is enough to start. Watch two clocks: the transaminases and the ALP plus bilirubin. Taper slowly. All of this comes from cohorts and expert consensus; there are no randomized trials.
Order and check
- ALT, AST, ALP, bilirubin, INR at least every 2 to 3 days
- Hepatitis B surface antigen, core antibody total, hepatitis C, CMV
- CK and troponin when AST leads
- Ultrasound with Doppler; MRCP if ALP leads
- GGT to confirm ALP comes from the liver
- Stop other liver drugs: acetaminophen, supplements, alcohol
- PJP prophylaxis: atovaquone if the liver is the problem
- Glucose, GI prophylaxis, calcium and vitamin D
Where people slip
- Watching only the transaminases. They can fall while the ALP and bilirubin climb. That is still steroid-refractory, and it is the pattern with the worst outcome.
- Grading by ALT alone. Use the higher of ALT and AST, and look at bilirubin. AST leading with high CK is muscle, maybe heart.
- Reaching for infliximab because it works in colitis. It is hepatotoxic; in the liver the second drug is mycophenolate.
- 2 mg/kg by reflex. Same liver outcome as 1 mg/kg, more infections and more hyperglycemia.
- Tapering in two weeks. One in four to one in three relapses on taper. Plan 4 to 6 weeks or longer, steroid off first, MMF last.
- Trimethoprim-sulfamethoxazole for PJP prophylaxis in a patient whose problem is the liver. Atovaquone avoids adding another hepatotoxic drug.
Two clocks. The transaminases tell you the steroid is working. The ALP and bilirubin tell you if it is not.
Sources
Retrieved from PubMed and the NCCN, ASCO and AGA guidelines, read on Nov 12, 2026.
- Schneider BJ, et al. Management of immune-related adverse events in patients treated with immune checkpoint inhibitor therapy: ASCO guideline update. J Clin Oncol 2021. doi:10.1200/JCO.21.01440
- Dougan M, et al. AGA clinical practice update on diagnosis and management of immune checkpoint inhibitor colitis and hepatitis. Gastroenterology 2021. doi:10.1053/j.gastro.2020.08.063
- NCCN Guidelines: Management of Immune Checkpoint Inhibitor-Related Toxicities, version 1.2026.
- Li M, et al. Effect of corticosteroid dosing on outcomes in high-grade immune checkpoint inhibitor hepatitis. Hepatology 2022. doi:10.1002/hep.32215
- Holmstroem RB, et al. Prospective study of patients with immune checkpoint inhibitor-induced hepatitis. J Immunother Cancer 2026. doi:10.1136/jitc-2025-013861
- Alouani E, et al. Mycophenolate mofetil for steroid-refractory immune-related hepatitis. Eur J Cancer 2023. doi:10.1016/j.ejca.2023.113313
- Romanski NA, et al. Risk factors and medical management of immune-related hepatotoxicity in metastatic melanoma. Eur J Cancer 2020. doi:10.1016/j.ejca.2020.02.041
Seminoma: AFP never, hCG sometimes, and the treatment follows the stage
Man in his 30s with a testicular mass. Pathology after orchiectomy: pure seminoma. AFP 80 ng/mL, hCG 12 IU/L. Seminoma or not?
Seminoma: germ cell tumor made only of seminoma cells; very sensitive to cisplatin and radiation. Nonseminoma: any tumor with embryonal, yolk sac, choriocarcinoma or teratoma parts, even if most is seminoma. AFP (alpha-fetoprotein) comes from yolk sac and embryonal cells. hCG (human chorionic gonadotropin) comes from syncytiotrophoblast cells, which a pure seminoma can contain. LDH: tumor bulk. IGCCCG risk: the international prognostic groups; a seminoma is never poor risk, only good or intermediate, and intermediate means metastases in an organ other than the lung (liver, bone, brain). BEP: bleomycin, etoposide, cisplatin. EP: etoposide, cisplatin. VIP: etoposide, ifosfamide, cisplatin. RPLND: retroperitoneal lymph node dissection. AUC 7: carboplatin dosed by kidney function to an area under the curve of 7. Gy: gray, radiation dose. PFS / OS: progression-free / overall survival. NPV / PPV: how often a negative / positive scan is right.
Algorithm
What the studies show
| Study | Who | Finding | Number to remember |
|---|---|---|---|
| Mortensen, Eur Urol 2014 | 1,954 stage I seminomas on surveillance, Denmark | Relapse in about 1 in 5 (19%), median time to relapse about 14 months; disease-specific survival 99.3% | 99%survive on surveillance |
| Oliver, J Clin Oncol 2011 (TE19) | 1,447 stage I, carboplatin x1 vs radiation | Carboplatin noninferior at 5 years, fewer second testis cancers | 94.7% vs 96.0%relapse-free, carboplatin vs radiation |
| Horwich, Br J Cancer 2000 | 130 metastatic seminomas, carboplatin alone vs EP | Carboplatin could not be shown equivalent; trial stopped early | 71% vs 81%3-year PFS, carboplatin vs EP |
| Beyer, J Clin Oncol 2021 (IGCCCG update) | 2,451 metastatic seminomas, 1990 to 2013 | LDH over 2.5 times normal is a new adverse factor within good risk | 95% vs 88%5-year OS, good vs intermediate |
| Bachner, Ann Oncol 2012 (SEMPET) | 127 PET scans of residual masses after chemotherapy | Timing matters: before 6 weeks, a positive PET is right only 1 in 4 times | 95% / 69%NPV / PPV at 6 weeks or later |
| Daneshmand, J Clin Oncol 2023 (SEMS) | 55 men, nodes 1 to 3 cm, surgery alone | 22% relapsed, and every relapse was salvaged with chemotherapy | 81%2-year relapse-free with surgery alone |
Bottom line: stage I seminoma is almost always cured whatever you choose, so pick the option with the least late toxicity. In metastatic disease, cisplatin is not negotiable.
Where people slip
- Accepting "pure seminoma" with a raised AFP. The pathologist sampled the wrong slice. Treat as nonseminoma. The one exception: a low, flat AFP from liver disease or a constitutional variant, which does not rise over serial checks.
- Giving carboplatin for metastatic disease because it works in stage I. Metastatic seminoma gets cisplatin.
- Irradiating bulky IIB. Radiation fits IIA and small, nonbulky IIB; bulkier nodes get chemotherapy.
- Ordering the PET too early. Before 6 weeks after chemotherapy, most positive scans are inflammation. And no PET for masses under 3 cm.
- Bleomycin by reflex. Get lung function and think of age, smoking, kidney function; EP x4 is the bleomycin-free option in good risk.
- Forgetting sperm banking before chemotherapy or radiation.
Seminoma makes hCG sometimes, AFP never. If AFP is up, it is not a pure seminoma.
Sources
Retrieved from PubMed, read on Nov 12, 2026.
- Mortensen MS, et al. A nationwide cohort study of stage I seminoma patients followed on a surveillance program. Eur Urol 2014. doi:10.1016/j.eururo.2014.07.001
- Oliver RT, et al. Randomized trial of carboplatin versus radiotherapy for stage I seminoma: mature results on relapse and contralateral testis cancer rates in MRC TE19/EORTC 30982 study. J Clin Oncol 2011. doi:10.1200/JCO.2009.26.4655
- Horwich A, et al. A Medical Research Council randomized trial of single agent carboplatin versus etoposide and cisplatin for advanced metastatic seminoma. Br J Cancer 2000. doi:10.1054/bjoc.2000.1498
- Beyer J, et al. Survival and new prognosticators in metastatic seminoma: results from the IGCCCG-Update Consortium. J Clin Oncol 2021. doi:10.1200/JCO.20.03292
- Bachner M, et al. FDG-PET for postchemotherapy seminoma residual lesions: a retrospective validation of the SEMPET trial. Ann Oncol 2012. doi:10.1093/annonc/mdr052
- Daneshmand S, et al. Surgery in early metastatic seminoma: a phase II trial of retroperitoneal lymph node dissection for testicular seminoma with limited retroperitoneal lymphadenopathy. J Clin Oncol 2023. doi:10.1200/JCO.22.00624
- NCCN Guidelines: Testicular Cancer, current version.
Inflammatory breast cancer, triple-negative and metastatic: how we treat it in 2026
Woman with a red, swollen, hot right breast that grew over weeks. ER 5%, PR negative, HER2 1+, germline BRCA2. Pleural effusion, supraclavicular nodes, liver lesions. What is first line?
Inflammatory breast cancer: a clinical diagnosis, not a histology: erythema and edema (peau d'orange) over at least a third of the breast, appearing in 6 months or less, with invasive carcinoma on biopsy; it is staged T4d (the T category reserved for inflammatory cancer). Triple-negative: ER and PR under 1%, HER2 0 to 1+ or 2+ with ISH negative. ER-low: ER 1 to 10%; behaves like triple-negative and is usually treated as such. IHC / ISH: immunohistochemistry / in situ hybridization. HER2-low: IHC 1+, or 2+ with ISH negative. BRCA1/2: homologous recombination genes; carriers respond to PARP inhibitors (olaparib, talazoparib) and platinum. CPS (combined positive score): PD-L1-stained tumor cells, lymphocytes and macrophages per 100 tumor cells, read with the 22C3 assay. ADC: antibody-drug conjugate. TROP2: the target of sacituzumab and Dato-DXd (datopotamab deruxtecan). T-DXd: trastuzumab deruxtecan. PFS: progression-free survival. OS: overall survival. Hazard ratio: under 1 favors the new drug.
Algorithm
What the studies show
| Study | Who | Finding | Number to remember |
|---|---|---|---|
| KEYNOTE-355, Cortes, N Engl J Med 2022 | 847 first line; chemotherapy with or without pembrolizumab | Benefit only when CPS is 10 or more | 23.0 vs 16.1months OS, CPS 10 or more, hazard ratio 0.73 |
| ASCENT-04, Tolaney, N Engl J Med 2026 | 443 first line, CPS 10 or more; sacituzumab vs chemotherapy, both with pembrolizumab | Longer PFS, longer responses, fewer stops for toxicity (12% vs 31%); OS immature | 11.2 vs 7.8months PFS, hazard ratio 0.65 |
| TROPION-Breast02, Dent, Ann Oncol 2026 | 644 first line, immunotherapy not an option; Dato-DXd vs chemotherapy | First ADC to improve OS in first line | 23.7 vs 18.7months OS, hazard ratio 0.79 |
| ASCENT-03, Cortés, N Engl J Med 2025 | 558 first line, not PD-1 candidates; sacituzumab vs chemotherapy | Longer PFS, same response rate; OS immature | 9.7 vs 6.9months PFS, hazard ratio 0.62 |
| ASCENT, Bardia, N Engl J Med 2021 | 468 after 2 or more lines; sacituzumab vs chemotherapy | Changed the later-line standard | 12.1 vs 6.7months OS, hazard ratio 0.48 |
| DESTINY-Breast04, Modi, N Engl J Med 2022 | 557 HER2-low; only 63 hormone receptor-negative (58 analyzed) | Benefit in the small triple-negative group, exploratory | 18.2 vs 8.3months OS, receptor-negative group |
| OlympiAD and EMBRACA, N Engl J Med 2017 and 2018 | Germline BRCA, HER2-negative, up to 2 (OlympiAD) or 3 (EMBRACA) prior lines; PARP inhibitor vs chemotherapy | Longer PFS; no significant OS gain overall | 0.58 / 0.54hazard ratio for PFS, olaparib / talazoparib |
| TNT, Tutt, Nat Med 2018 | 376 first line; carboplatin vs docetaxel | Carboplatin doubled response only in germline BRCA carriers | 68% vs 33%response in BRCA carriers |
Bottom line: in 2026 first line is an ADC for almost everyone. The two questions are CPS and whether the patient can receive PD-1. Overall survival benefit from an ADC in first line is shown so far only for Dato-DXd.
Where people slip
- Treating ER-low as hormone receptor-positive and giving endocrine therapy with a CDK4/6 inhibitor (palbociclib, ribociclib, abemaciclib). It is triple-negative biology.
- Reading HER2 and PD-L1 on old tissue only. Biopsy the metastasis: HER2-low and CPS can change, and they decide the drug.
- Using the wrong PD-L1 assay. Pembrolizumab decisions use CPS by 22C3; SP142 belongs to atezolizumab.
- Forgetting the neoadjuvant pembrolizumab. After KEYNOTE-522 most patients already had PD-1 in the curative setting: that puts them on the not-a-candidate side. And relapse within 6 months of any curative therapy was excluded from every first-line trial: there are no first-line data, so treat it as second line.
- Expecting a second ADC to work as well as the first. Sacituzumab, Dato-DXd and T-DXd all carry a topoisomerase I payload; there are no randomized sequencing data.
- Mastectomy up front in stage IV inflammatory disease. Systemic therapy first; radiation for local control.
Two questions, then the ADC. CPS 10 or more? PD-1 candidate? Yes to both: sacituzumab with pembrolizumab. No: Dato-DXd or sacituzumab alone. BRCA: keep the PARP inhibitor and the platinum in the back pocket.
Sources
Retrieved from PubMed and FDA approvals, read on Nov 12, 2026.
- Dawood S, et al. International expert panel on inflammatory breast cancer: consensus statement for standardized diagnosis and treatment. Ann Oncol 2011. doi:10.1093/annonc/mdq345
- Allison KH, et al. Estrogen and progesterone receptor testing in breast cancer: ASCO/CAP guideline update. J Clin Oncol 2020. doi:10.1200/JCO.19.02309
- Cortes J, et al. Pembrolizumab plus chemotherapy in advanced triple-negative breast cancer. N Engl J Med 2022. doi:10.1056/NEJMoa2202809
- Tolaney SM, et al. Sacituzumab govitecan plus pembrolizumab for advanced triple-negative breast cancer. N Engl J Med 2026. doi:10.1056/NEJMoa2508959
- Dent R, et al. Datopotamab deruxtecan in patients with untreated, advanced triple-negative breast cancer (TROPION-Breast02). Ann Oncol 2026. doi:10.1016/j.annonc.2026.03.008
- Cortés J, et al. Sacituzumab govitecan in untreated, advanced triple-negative breast cancer. N Engl J Med 2025. doi:10.1056/NEJMoa2511734
- Bardia A, et al. Sacituzumab govitecan in metastatic triple-negative breast cancer. N Engl J Med 2021. doi:10.1056/NEJMoa2028485
- Modi S, et al. Trastuzumab deruxtecan in previously treated HER2-low advanced breast cancer. N Engl J Med 2022. doi:10.1056/NEJMoa2203690
- Robson M, et al. Olaparib for metastatic breast cancer in patients with a germline BRCA mutation. N Engl J Med 2017. doi:10.1056/NEJMoa1706450
- Litton JK, et al. Talazoparib in patients with advanced breast cancer and a germline BRCA mutation. N Engl J Med 2018. doi:10.1056/NEJMoa1802905
- Tutt A, et al. Carboplatin in BRCA1/2-mutated and triple-negative breast cancer BRCAness subgroups: the TNT trial. Nat Med 2018. doi:10.1038/s41591-018-0009-7
- FDA approvals: datopotamab deruxtecan, May 22, 2026; sacituzumab govitecan alone and with pembrolizumab, June 24, 2026, first-line metastatic triple-negative breast cancer.