Correcting induced false-positive events and using a new data interpretation algorithm may make droplet digital PCR (ddPCR) more sensitive and accurate in patients with stage 4 non-small cell lung cancer (NSCLC) (Clin Chem 2021; doi:10.1093/clinchem/hvab040).

Often used to detect mutations in cancer patients’ circulating cell-free (cf) DNA, ddPCR is highly sensitive and can detect variants with low variant allele frequencies. Based on water-oil emulsion droplet technology, ddPCR involves fractionating the sample into thousands of droplets and amplifying template molecules in each individual droplet.

Distinguishing a true positive from a false positive signal is a challenge with ddPCR. While optimum sensitivity and specificity rely on correct interpretation of ddPCR results, there is no standardized method for interpreting ddPCR data. Meanwhile, few studies have focused on technical artifacts that affect ddPCR results. 

In response, researchers aimed to report the occurrence of polymerase-induced false-positive events (PIFs) and an input-dependent increase in PIFs in ddPCR experiments. The researchers developed a novel ddPCR data interpretation algorithm called “adaptive limit of blank (LoB) and PIFs: an automated correction algorithm” (ALPACA). The algorithm combines corrections for assay-specific error rates and technical artifacts.

The researchers determined false positive rates for six ddPCR assays at varying amounts of input DNA. This process revealed PIFs and other false positives. The researchers used an in silico correction algorithm plus ALPACA to remove PIFs and to apply an adaptive LoB to each sample. Afterwards, the researchers compared ALPACA’s performance to a standard strategy that did not involve PIF correction, using a LoB of 3 and data from commercial reference DNA. The researchers conducted this comparison on cfDNA from healthy volunteers and a cohort of 209 patients with stage 4 NSCLC, as well as their molecularly profiled tumors.

Applying ALPACA reduced false-positive results in cfDNA of healthy volunteers and NSCLC patients, compared to the standard strategy. For healthy volunteers, specificity was 98% using ALPACA, versus 88% for the standard method. For stage 4 NSCLC patients, specificity was 99%, versus 93% using the standard method. Using ALPACA did not markedly affect sensitivity in commercial reference DNA or patient cfDNA.


A recent paper makes a case for preconception thyrotropin screening (JAMA Network Open 2021;4(4):e215723. doi:10.1001/jamanetworkopen.2021.5723). 

Maternal thyroid function—especially during early gestation—is important for the best pregnancy outcomes. Previous research suggests an early, narrow window for diagnosing and treating thyroid maladies. Although medical guidelines recommend screening for thyroid dysfunction during pregnancy, little research has examined preconception screening.

In a population-based cohort study of approximately 5.8 million Chinese women ages 20 to 49, researchers studied associations between preconception thyrotropin levels and preterm birth (PTB), small size for gestational age (SGA), birth defects, and perinatal infant death. The mothers gave blood after at least 8 hours of fasting prior to preconception exams conducted within 6 months before pregnancy. The women also had exams in early pregnancy and following adverse outcomes.

The researchers established a population-specific reference range that established the 2.5th, 50th, and 97.5th percentiles for thyrotropin levels as 0.37 mIU/L, 1.66 mIU/L, and 4.88 mIU/L, respectively.

The women’s median thyrotropin level was 1.60 mIU/L. Cumulative incidences for adverse pregnancy outcomes were 6.56% for PTB, 7.21% for SGA, 0.02% for birth defects, and 0.33% for perinatal infant death. Compared with a reference group with thyrotropin range concentrations ranging from 0.37-2.49 mIU/L, both low (less than 0.10 mIU/L  to 0.36 mIU/L) and high (4.88 to 10.00 mIU/L) maternal preconception thyrotropin levels were associated with higher risk of PTB, SGA, and perinatal infant death. The researchers found J-shaped associations—indicating lowest incidence in the middle ranges and peaks in the upper and lower ranges—between preconception thyrotropin levels and PTB and perinatal infant death.

Based on these findings, the researchers suggest that an optimal preconception thyrotropin level is between 0.37 mIU/L and 2.50 mIU/L.

The researchers note some study limitations. These include not testing thyroid autoantibodies, which are associated with a higher risk of PTB, not following up on offspring’s intellectual and cognitive function, and stratifying patients according to iodine status.


Serial pancreatic stone protein (PSP) increases in the days preceding the onset of signs involved in clinical diagnosis of sepsis, according to recent research (Crit Care 2021:25;151).

Biomarkers may help identify subclinical signs of sepsis, leading to earlier recognition and management of sepsis and better outcomes. Researchers conducted an observational clinical study to explore how well serial measurements of C-reactive protein (CRP), procalcitonin (PCT), and pancreatic stone protein (PSP) would spot sepsis in 234 patients at risk of nosocomial sepsis in 14 intensive care units (ICUs) in France, Switzerland, Italy, and the United Kingdom.

Via receiver operating characteristic (ROC) analysis, the researchers assessed these biomarkers’ performance as sepsis biomarkers by examining the association of clinical diagnoses with the trajectories of PSP, CRP, and PCT during the 3 days preceding diagnosis. In the absence of an unambiguous definition of sepsis and highly accurate diagnostic tools, an expert committee determined the presence or absence of sepsis on any given day of an ICU stay.

Fifty-three patients developed nosocomial sepsis after a median of 6 days. The diagnostic accuracy of PSP, CRP, and PCT were similar. Sepsis diagnosis was associated with an increase in all biomarkers’ value over the 3 days preceding this. However, PSP started to increase 5 days before the clinical diagnosis of sepsis, compared to 3 days prior for PCT, and 2 days prior for CRP. The area under the ROC curve at the time of clinical sepsis was similar for all markers (PSP, 0.75; CRP, 0.77; and PCT, 0.75).

While the diagnostic accuracy of PSP, CRP, and PCT for sepsis were similar, the rise in PSP before the others justifies further evaluation of it as a serially measured biomarker for managing critically ill patients.