Tracking Residual Lung Cancer With a Drop of Blood, While Cord Blood Cell Composition Shapes CAR-NK Efficacy

Whether cancer treatment works sometimes depends not only on which drug is used, but also on changes in the blood before and after treatment and the original composition of the cell product. Two studies recently highlighted by MD Anderson Cancer Center identified important clues that may affect the assessment of treatment efficacy in resectable non-small cell lung cancer and CAR-NK cell therapy, respectively.

Published in *Nature*, the CheckMate 77T biomarker analysis included 190 evaluable patients. This phase 3 trial compared a perioperative regimen of nivolumab plus chemotherapy before surgery followed by nivolumab after surgery with the same chemotherapy combined with placebo. One focus of the study was whether circulating tumor DNA (ctDNA) in the blood could be cleared before surgery.

Among patients with evaluable ctDNA both before and after treatment, 66% in the nivolumab group achieved ctDNA clearance before surgery, compared with 38% in the placebo group. Among those with ctDNA clearance, the pathological complete response rates were 50% and 12%, respectively, indicating that the disappearance of tumor signals from the blood was associated with the absence of viable cancer cells in resected specimens, although the two measures are not interchangeable.

Postoperative monitoring produced an even clearer warning signal: all 13 patients who were initially negative for minimal residual disease (MRD) but became positive during adjuvant treatment subsequently experienced recurrence. Machine-learning analysis also identified preoperative ctDNA clearance, absence of N2 lymph node involvement, pathological complete response, squamous histology, and nivolumab treatment as the main predictors of longer event-free survival. However, the biomarker analysis covered only a subset of all randomized patients, and the model was developed using the same evaluable dataset. The findings therefore remain exploratory and are not yet sufficient on their own to determine whether treatment should be stopped or extended, or whether follow-up practices should be changed.

Another study, published in *Cancer Cell*, shifted the focus to the starting point of CAR-NK manufacturing. The research team found that a population of immature NK cells in cord blood lacking CD16 and CD161 was associated with poorer CAR-NK treatment performance. Cord blood units containing a higher proportion of mature NK cells, meanwhile, were linked to stronger responses and better patient outcomes.

Mechanistic studies showed that these immature cells acquire proteins from the surface of tumor cells through “trogocytosis,” thereby becoming false targets carrying tumor antigens. Engineered CAR-NK cells may not only mistakenly attack their peers but also undergo activation-induced cell death and exhaustion. The immature cells also act like an “antigen sink,” continually consuming the effector cells that possess genuine killing capacity.

After researchers removed this cell population before CAR engineering, they observed better cell persistence, tumor control, and survival in preclinical models of lymphoma and ovarian cancer. This suggests a specific process-improvement pathway for off-the-shelf CAR-NK products: first analyze the cell subpopulations in donated cord blood and then remove them as needed. However, the improvements currently come mainly from model experiments and still need to be confirmed in prospective clinical studies. It will also be necessary to assess whether the additional screening step can maintain consistency in large-scale manufacturing.