The Leukocyte ImmunoTest for identifying radiorecurrent prostate cancer: a pilot study

Inflammation is a cancer hallmark, with evidence suggesting neutrophils contribute to the initiation, development, and progression of many cancers [1]. Neutrophil-to-lymphocyte ratio, a marker of neutrophil number, reflects prognosis in prostate cancer [2]; markers of neutrophil function could have a comparable role [2]. This may be important post-radiotherapy. >12,000 UK men undergo external beam radiotherapy (EBRT) yearly, but approximately 20% develop biochemical recurrence within 5 years [3,4]. Typically, PSA rises trigger re-staging, with the Phoenix threshold of ≥2.0ng/mL above-nadir often used [5]. A simple, accurate test alongside PSA may clarify suspicion of recurrence at a given timepoint and could facilitate earlier re-staging.

We performed a pilot study for this purpose, evaluating the novel Leukocyte ImmunoTest^™ (LIT; Seroxo Ltd, UK). This point-of-care assay assesses leucocyte ability to produce reactive oxygen species. We hypothesised higher LIT scores, reflecting increased neutrophil activity, may correspond with the presence and degree of radiorecurrence. Participants were part of the FORECAST prospective paired-cohort study (NCT01883128) [6]. Between 2014–2018, 181 patients were enrolled from 6 UK centres with biochemical failure after EBRT or brachytherapy with/without (neo)adjuvant androgen deprivation therapy (ADT). Biochemical failure was defined by the referring clinician; our protocol did not mandate the Phoenix criteria. After having blood taken for LIT, patients were re-staged with ¹⁸F-Choline PET/CT and ^99mTc-MDP bone scan before multiparametric MRI. All patients then underwent transperineal template mapping biopsies. Patients with MRI lesions of Likert score 3–5 also underwent cognitive MRI-targeted biopsy.

The LIT assay has been previously described [7]. A finger-prick blood sample was added to phosphate-buffered saline containing phorbol-12-myrisate-13-acetate. This stresses leucocytes to release reactive oxygen species, which react with luminol, causing photon emission. After incubation for 10 minutes at 37.5°C, chemiluminescence was measured in relative light units (RLU) using a handheld luminometer. Patients underwent the assay thrice, with scores averaged.

Jitter plots, Kruskal-Wallis testing, and univariable ordinal logistic regression were used to examine the association between mean LIT score (mLIT) and enrolment PSA with two ordinal outcomes. The first was disease stage: no cancer, localised cancer (N0M0), and nodal/metastatic cancer (N1 and/or M1). The second was pathological characteristics: no cancer, grade group (GG) 1–2 with maximum cancer core length (MCCL) 1–5mm, and GG 3–5 or GG 1–2 with MCCL ≥6mm (PROMIS definition 1). For this, we acknowledge the precise role of Gleason grading in radiorecurrence has not been established. Furthermore, these definitions reflect a spectrum of pathological outcomes aligning to definitions used in the primary setting, rather than groupings explicitly corresponding to different treatment recommendations. Missing data for enrolment PSA in 3/70 men (4%) were derived using multiple imputation by chain equations with 20 iterations and 500 resamples. mLIT and PSA were log-transformed for regression modelling.

70 patients consented to LIT and completed all investigations, thus were included (Fig.S1). At enrolment, there was a median of 7.5 years from original diagnosis (IQR 5.2–11.0). Median enrolment PSA was 4.6ng/mL (IQR 3.1–6.9) (Table S1). Median PSA increase-from-nadir, calculable in 67 patients with complete data, was 4.5ng/mL (IQR 2.8–8.3), with 60/67 patients (90%) meeting Phoenix criteria. 65/70 patients (93%) received EBRT, and 6/70 (9%) brachytherapy. 33/70 patients (47%) received neoadjuvant ADT only, and 13/70 (19%) adjuvant ADT. Median mLIT score was 296RLU (IQR 194–470, range 75–1782). 65/70 patients (93%) had recurrent cancer identified on prostate biopsy alone (48/70; 69%), as extra-prostatic disease on imaging with negative prostate biopsy (3/70; 4%), or cancer identified both on prostate biopsy and as extra-prostatic disease on imaging (14/70; 20%). Of the 5 patients with no cancer identified, 3 met Phoenix criteria, 1 did not, and 1 did not have a calculable PSA increase-from-nadir. There was no statistically significant association between radiorecurrent cancer with mLIT (p=0.7) nor enrolment PSA (p=0.8).

Fig.1 and Table S2 display mLIT and enrolment PSA categorised by disease stage. On inspection, mLIT distribution appeared comparable across stages though the nodal/metastatic group did contain the four patients with the highest values. There was no statistically significant difference in mLIT between stages (p=0.5). There was also no statistically significant association between 50RLU increments in mLIT and stages (OR 1.31, 95%CI 0.58–2.97, p=0.5). On inspection of PSA distribution, patients with higher values were more concentrated to localised and nodal/metastatic disease groups. There was no statistically significance difference in enrolment PSA between patients across stages (p=0.6). In regression, there was no statistically significant association between 1.0ng/mL increments in enrolment PSA and stages (OR 0.98, 95%CI 0.37–2.63, p=0.97).

Fig.1:

Jitter plots showing the distribution of mLIT and enrolment PSA with disease stage (A and B, respectively) and pathological characteristics ordinal outcomes (C and D, respectively).

Pathological characteristics was studied next. On inspection of mLIT distribution, this appeared comparable between the three pathological groups (Fig.1). However, the four men with the highest scores were in the PROMIS definition 1 group. There was no statistically significant difference in mLIT between pathological groups (p=0.6), nor any statistically significant association between 50RLU increments and pathological grouping (OR 1.24, 95%CI 0.50–3.09, p=0.6). Enrolment PSA had a comparable distribution to mLIT. There was also no statistically significant difference in PSA between pathological groupings (p=0.8), and no statistically significant association between 1.0ng/mL increments and pathological groupings in regression (OR 1.82, 95%CI 0.56–5.94, p=0.3).

Use of the LIT assay was feasible alongside PSA surveillance. In this small cohort, no statistically significant associations were detected between mLIT or PSA with two ordinal outcomes of stage and pathological characteristics. However, on jitter plot inspection, there was a suggestion that the highest mLIT scores are observed in patients with nodal/metastatic cancer and with the highest grade and/or volume disease. Further study in larger, powered cohorts is warranted to evaluate this finding. Given this is a point-of-care blood test with results available in 10 minutes, the LIT assay could have value in assessing men with a rising PSA post-radiotherapy. Potentially, a high LIT score with a rising PSA could prompt earlier triage towards re-imaging, rather than further PSA surveillance.

Earlier detection of radiorecurrence is important. Within 5 years of biochemical failure, approximately 50% develop metastases and 20% experience cancer-specific death [4]. However, interpretation of PSA post-radiotherapy can be difficult owing to influences like the bounce phenomenon, testosterone recovery post-ADT, sexual activity, and infection. Accordingly, PSA thresholds can be limited by false positives and negatives [5,8]. A rising PSA also does not easily distinguish between local recurrence, regional spread, and distant metastases. An adjunctive, accurate test could therefore be useful for identifying recurrence and degree-of-spread.

Study of LIT values in surveillance prior to the point of PSA biochemical failure warrants further investigation. Alongside further testing in larger cohorts with paired LIT and PSA surveillance, decision-making and cost-effectiveness need additional evaluation. Ideally, development of a nomogram using mLIT, PSA, and other variables could prompt earlier detection of radiorecurrence. We also acknowledge a study limitation is use of ¹⁸F-Choline PET/CT and bone scan for staging, the standard at the time. A future consideration is whether PSMA PET/CT may better detect occult metastases. Robust verification of extra-prostatic imaging findings, whether by biopsy or clinical measures, should also be considered.

Our pilot study shows that LIT, a simple point-of-care blood test, is feasible for investigating patients with a rising PSA post-radiotherapy. There was a suggestion that the highest LIT scores were recorded in patients with extra-prostatic spread and patients with higher grade and/or volume disease on prostate biopsy. Evaluation of this observation in larger cohorts and with PSMA PET/CT is now needed.