It has been our practice to send a sample of cerebrospinal fluid (CSF) for cytospin at the time of each lumbar puncture (LP) in children undergoing treatment for acute lymphoblastic leukaemia (ALL) in all the trials and treatment guidelines over the last 20 years at Alder Hey Children’s Hospital.

A recent study has looked at the role of CSF surveillance in detecting silent CNS relapse in 331 children enrolled on UKALL 2003 who had completed treatment by May 2011 at Great Ormond Street and University College Hospital, London. The number of asymptomatic children with a positive CSF cytospin was divided by the total number of cytospins in the cohort revealing a detection rate of 0.09% (Samarasinghe et al 2012).

In this respect we undertook a retrospective study at our institution to further examine the benefit of routine CSF cytology in patients with ALL. In particular, we aimed to determine the utility of CSF surveillance in detecting asymptomatic CNS relapse on treatment.


The medical records of all children diagnosed with ALL from 1992 onwards, and who had completed treatment by May 2013 were examined.

CNS relapse was diagnosed if blasts were detectable on CSF cytocentrifuge (>5/µL). Relapses were classified as symptomatic if they had signs suggestive of CNS leukaemia such as headache, diplopia or cranial nerve palsies and asymptomatic if such signs and symptoms were absent.


407 eligible patients were identified. There were 224 males and 183 females, with a median age of 6.5 years (range 2 months to 19 years).

The subtype of ALL by immunophenotyping wad B lineage in 83%, T lineage in 12.25%, Philadelphia positive in 0.5%, null cell in 0.25% and subtype unknown in 4%.

The cohort received treatment under a number of different UK protocols throughout the last 21 years as follows: 81 patients UKALL XI, 6 patients UKALL XI HR, 39 patients ALL 97, 9 patients ALL 97 HR, 83 patients ALL 97/99, 17 patients ALL 2003 MRD Pilot, 157 patients ALL 2003, 13 patients Interfant, 2 patients EsphALL.

There were 90 relapses in 407 patients; 38 of these occurring in patients whilst on treatment.

Seven out of these 38 relapses were CNS relapses – 5 isolated CNS relapse and 2 combined bone marrow and CNS relapse. Of these seven patients with CNS relapse on therapy, six had clear symptoms and signs of CNS relapse including headache (5 patients), vomiting (2 patients), and cranial nerve palsy (4 patients). Such symptoms or associated bone marrow disease would have prompted CSF cytology.

Only one patient was diagnosed with asymptomatic CNS relapse on routine CSF examination, but this patient became symptomatic the following day with a facial nerve palsy.

Six of these 7 patients with CNS relapse have since died; 5 following leukaemic relapse and 1 following lateral sinus thrombosis.


This retrospective study confirms the very low rate of detection of asymptomatic CNS relapse for patients receiving treatment for ALL. The benefit of routine CSF cytology i.e. at the time of intrathecal chemotherapy is thus very much drawn into question.

We have thus changed practice and only undertake routine CSF cytology in the following situations: (1) at diagnosis in all patients, (2) throughout treatment in patients with CNS disease at diagnosis, (3) throughout treatment in patients treated on protocols for infant ALL and Philadelphia positive ALL.

We do, however, recognise the need for a low threshold for performing an LP in any patient who shows signs or symptoms of CNS leukaemia.

In summary, there appears to be little value in routine surveillance of CSF in the large majority of children and young people with acute lymphoblastic leukaemia.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.