A standard of care for individuals with PIK3CA‐related disorders: An international expert consensus statement

Growth promoting variants in PIK3CA cause a spectrum of developmental disorders, depending on the developmental timing of the mutation and tissues involved. These phenotypically heterogeneous entities have been grouped as PIK3CA‐Related Overgrowth Spectrum disorders (PROS). Deep sequencing technologies have facilitated detection of low‐level mosaic, often necessitating testing of tissues other than blood. Since clinical management practices vary considerably among healthcare professionals and services across different countries, a consensus on management guidelines is needed. Clinical heterogeneity within this spectrum leads to challenges in establishing management recommendations, which must be based on patient‐specific considerations. Moreover, as most of these conditions are rare, affected families may lack access to the medical expertise that is needed to help address the multi‐system and often complex medical issues seen with PROS. In March 2019, macrocephaly‐capillary malformation (M‐CM) patient organizations hosted an expert meeting in Manchester, United Kingdom, to help address these challenges with regards to M‐CM syndrome. We have expanded the scope of this project to cover PROS and developed this consensus statement on the preferred approach for managing affected individuals based on our current knowledge.


| INTRODUCTION
Activating variants in PIK3CA cause a phenotypic spectrum of developmental disorders known as the PIK3CA-Related Overgrowth Spectrum (PROS) (Table 1(A)). [1][2][3][4][5][6][7][8][9][10] Deep sequencing technologies have facilitated the identification and clinical testing for PROS-related gain-of-function (GoF) variants that are often either absent or present at very low levels, in peripheral blood. Typically, testing other types of tissues or biopsies is necessary for detection of ultra-low frequency mosaic variants. [11][12][13][14] As variants are typically growth-promoting, a random and post-zygotic mutational event in PIK3CA may provide the cell of origin with a selective growth advantage. The phenotypic consequences are dependent on when, during development, and where (in which cell) the mutational event occurred, on the distribution of the mutated cell's progeny as well as on the activating strength of the variant.
These mosaic disorders are quite variable in nature, extent, progression, and severity of the clinical presentation. Conversely, deep phenotyping has shown that despite marked diversity, these conditions are also clinically overlapping. 18,19 While many affected individuals have all manifestations of, and are typical for, previously named entities (e.g., MCAP or CLOVES), there are innumerable individuals who manifest only some aspects of these entities and those who manifest attributes of more than one of these entities but within the broad spectrum of PROS. Recently, we have proposed a dyadic approach to the diagnosis and terminology of entities and we believe this may be equally applicable for PROS. 20 This extensive phenotypic overlap warrants a holistic, multidisciplinary approach with expert-driven medical management, and taking into considerations available clinical screening tools in various healthcare settings. Since clinical management practices vary considerably among different professionals, healthcare services, and countries, general management guidelines are needed. [21][22][23] Moreover, as most of these conditions are rare, affected families may lack access to the medical expertise that is needed to help address the multi-system and often complex medical issues seen with PROS. An international collaboration to derive evidence-based management and surveillance guidelines was stimulated and funded by M-CM/MCAP patientadvocacy and support groups from the United Kingdom, the United States and Spain. 24

| MATERIALS AND METHODS
These are detailed in Figure 1. The information contained in the publications chosen following literature curation was sent to a panel of experts and assessed using the AGREE II tool. 25 We asked each expert to derive management recommendations based on the evidence provided but also on further evidence from their own field if additional recommendations were necessary. The experts were allowed to abstain from scoring any recommendation if it was outside their area of expertise. Recommendations were collated and discussed at the expert consensus meeting held in Manchester, U K, in March 2019.
Following the expert consensus meeting, a first draft of the consensus statement was circulated to the group for review. The final, consensus document is presented here. The majority of the recommendations presented here scored '+++' for Agreement (>75% agree with recommendation) and 'A' for Evidence (based on evidence +/À expert consensus) (see the AGREE II tool, Figure 1). We list instead, in Table 6, the recommendations with lower scores.

| Diagnostic criteria and nomenclature
The PROS umbrella term and diagnostic criteria were defined by Keppler-Noreuil et al. Table 1(B) exemplifies the distinction we developed to guide the clinical management of these entities, once molecularly confirmed. We have defined two broad phenotypic categories to which these management guidelines should be applied: MCAP and non-MCAP PROS. When one of these terms is used hereafter, we intend that the recommendation applies to that category. If we instead specify ''PROS'' the recommendation applies to both categories.
It is important to note that we are not suggesting that these descriptors supplant specific clinical terms such as CLOVES, rather, it is a heuristic categorization that we believe is useful to classify individuals into groups for application of our management recommendations. We appreciate that families and healthcare providers are concerned about changes in nomenclature, but the insights in disease pathogenesis obtained from progress in genomics, pathology, and neuroradiology require regular adaptations. Irrespective of any of our preferences, the nearly continuous inter-patient variation of PROS, especially that associated with mosaic alterations, forces us to adopt pragmatic approaches and we acknowledge that these approaches may and will evolve with time.

| General discussion
Establishing a molecular diagnosis in PROS is important and applying management strategies to individuals without a molecular diagnosis can be very challenging. A molecular confirmation may guide patient management and monitoring, offer new therapeutic approaches, where available, and allow provision of appropriate recurrence risks.
Testing DNA from affected tissue with highly sensitive diagnostic methods offers the best chance of detecting postzygotic/somatic variants in PIK3CA. 5,12,13 Cell-free DNA (cfDNA) is now routinely used in cancer applications including testing urine samples in patients with PROS and kidney involvement. 26 However, the diagnostic yield may not be as high as in cancer applications as: (a) the probability of cfDNA carrying the mutation may vary with clinical aspects (i.e., presence of a vascular malformation). In PROS, there is probably not as much or any neovascularization as in cancer; (b) the VAF with cfDNA can be very low, and ddPCR is used instead of NGS. ddPCR can only screen about five variants and is thus limited to hotspots screening (or confirmation of NGS screening). Taking into consideration the allelic heterogeneity of PROS, 13 false negative tests are expected; (c) it is difficult to interpret the biological significance of very low VAFs (0.001%). Overall, cfDNA testing can be compared, at the moment, with blood or saliva testing against affected tissue testing: it is less invasive, but the diagnostic yield is lower. Further systematic data are needed to advance any recommendations about its use in a diagnostic setting.

Recommendations for MCAP
It is possible to detect pathogenic variants in peripheral blood or saliva samples in some individuals. The level of mosaicism in blood or saliva samples is usually low; thus, the diagnostic yield is lower 5,13 (see also Appendix).

Recommendations for non-MCAP PROS
In highly focal disorders within the PROS spectrum (e.g., CLOVES, fibroadipose hyperplasia), pathogenic variants are detectable in affected tissues with wide ranges of mosaicism. There is conflicting evidence about correlation between diagnostic rate and localized or tissue-specific presentations. 5,9,12,13 This problem, however, could be restricted to certain cases of PROS, e.g., PIK3CA-related muscular overgrowth, and due to the very restricted distribution of affected tissue. In such cases, NGS following sampling of affected tissue is the best option (e.g., during reduction surgery of macrodactyly or for debulking of ectopic accessory muscles).

| Recommendations for MCAP
Individuals with MCAP may have complex feeding difficulties that should be assessed and taken into consideration in monitoring weight gain.
T A B L E 2 Genetic testing recommendations for PROS

Method
The molecular diagnostic approach should utilize techniques optimized for the detection of low-level mosaic variants (low variant allele fraction or VAF, see also Appendix). Because many of the variants that cause PROS are typically absent in peripheral blood, Next Generation Sequencing (NGS) or digital droplet polymerase chain reaction (ddPCR) testing Sampling Affected tissues (i.e., during reduction surgery of macrodactyly or for debulking of ectopic accessory muscles) offer the optimal approach. When tissue samples are not available, a biopsy of affected tissue (e.g., skin with a vascular malformation in a region with soft tissue overgrowth) may be considered. DNA should be extracted without previous culture. Genetic testing on formalin-fixed, paraffin-embedded (FFPE) samples from prior surgeries is also possible. Although, in this type of sample, the DNA may be degraded, reducing the diagnostic yield Optimal depth of NGS The absolute minimum number of times that a variant must be determined depends on the platform used and its ability to discriminate against background noise. For example, on hybridization-based NGS platforms, a minimum 350x and mean 500x coverage may be sufficient to detect VAFs of 5%. These numbers can be considerably higher on amplicon-based NGS platforms

Detection of low-level mosaicism
Previous studies have shown a lower diagnostic rate in individuals presenting with very localized or tissue-specific features (e.g., isolated macrodactyly). 5 In a clinical context, diagnostic techniques able to detect mosaic variants with allele frequencies or VAF as low as 5% (0.05), or even lower at 1% (0.01) should be selected

Variant validation
Validation by orthogonal methods (e.g., Sanger sequencing for VAFs ≥20%, pyrosequencing for VAFs ≥5%, or ddPCR for VAFs ≤5%) or by a second NGS study of mosaic candidate variants may be needed. Variant validation might not be necessary when the candidate variant is detected in more than one different tissue of the same patient (taking into consideration the limit of detection of the technique). When validating variants in PIK3CA, it is necessary to remember that exons 10 to 14 of this gene share 95% identity with another region or pseudogene of the genome located on chromosome 22. Therefore, validation tests must be designed to discriminate the location of these variants (e.g., using long-range PCR upstream of the final PCR amplification) 7 Approach to negative testing results Given that low-level mosaicism is observed in many individuals with PROS, it is recommended to re-evaluate the quality of the sample (e.g., FFPE samples), the sampled tissue, the mosaic detection limit of the technique, and the clinical diagnosis

| Recommendations for PROS
A cranial MRI scan should be performed in all individuals with PROS in the presence of (a) rapidly enlarging occipital-frontal circumference (OFC) and/or (b) moderate-to-severe developmental delay or intellectual disability (ID) (c) epilepsy or (d) facial or craniofacial neurological involvement. In all affected individuals with brain involvement there is potential for hydrocephalus which can be due to venous Recommendations for regular monitoring of diagnosed CNS overgrowth or dysplasia From birth until age 8 years Cranial MRI every 6 months until age 2 and then annually until age 8 years to monitor specifically for progressive ventriculomegaly/hydrocephalus and CBTE or Chiari malformations. Brain MRIs may necessitate sedation in children and the risk to benefit ratio need to be weighed on an individual basis. An exception can be made for children over 4 years of age who are asymptomatic or have no signs suggesting raised ICP or Chiari and whose OFCs are more or less parallel to the standard growth curve (albeit large). In such instances, we recommend yearly follow-up with a pediatric neurologist for routine monitoring.
If an individual is diagnosed between 2 and 8 years of age A baseline brain MRI is recommended at the time of diagnosis with yearly MRI scans to monitor for progressive ventriculomegaly/hydrocephaly and CBTE. Additionally, a brain MRI should be considered if the head growth trajectory rapidly accelerates (for e.g., with crossing of OFC centiles on the growth chart).
Children with ventriculomegaly or CBTE Should be referred for neurosurgical evaluation, management, and monitoring for progression. The decision for neurosurgical intervention, including placement of a ventriculoperitoneal shunt, endoscopic third ventriculostomy or posterior fossa decompression, requires careful consideration by a genetics/pediatric neurosurgical team and is beyond the scope of the current recommendations.
Management of CBTE Nonsurgical management for stable, asymptomatic CBTE is appropriate, and decompression should be considered on a case-by-case basis especially when symptoms of brainstem compression appear or syringomyelia develops. This requires the input of a multi-disciplinary team including a genetics/pediatric neurosurgical team familiar with the care of this complex syndrome. hypertension/congestion or cerebellar tonsillar herniation (CBTE). centile) at birth that may rapidly progress to OFCs above two standard deviations or more, usually within the first 1-2 years of life.

B. CNS critical care and mortality risk recommendations
Progressive ventriculomegaly and hydrocephalus in these children is multifactorial and may be a result of both obstructive and communicating hydrocephalus, especially when the child develops cerebellar tonsillar herniation with posterior fossa cerebellar overgrowth. We caution the use of ventriculoperitoneal shunting in an asymptomatic child without progressive ventriculomegaly. We recommend that a pediatric neurosurgeon familiar with this condition is involved in the decision making and management of ventriculomegaly in these children (see also Epilepsy may be a significant concern for individuals with PROS with brain involvement. Cortical malformations seen in individuals with PROS may affect the severity and frequency of epilepsy. For example, epilepsy associated with PMG (not specifically PROSrelated) is often focal in origin, though it can also have generalized onset. 38 The precise incidence of epilepsy is not known but has been estimated to be as high as 50% in children with bilateral cortical malformations. 38 It is important to raise awareness among parents and caregivers that episodic irritability, loss of tone, loss of consciousness, or seizures may be wholly attributable to, or exacerbated by, hypoglycaemia. These are most likely to occur as feeding intervals increase, or during intercurrent illness or other stress

Baseline screening
Baseline screening of thyroid function (critically including both TSH and free thyroxine determination) in all patients with MCAP or other forms of PROS with brain involvement (see Table 1)

Endocrine supervision
Symptoms suggestive of periodic hypoglycaemia should trigger an inpatient fast under endocrine supervision. Attention should be paid to obtaining the correct diagnostic samples at the time of symptoms. Confirmed hypoglycaemia should be followed up by screening of GH and cortisol secretion

Monitoring of linear growth
Where there is evidence of linear growth retardation, IGF1 and IGFBP3 should be measured. If these are low, then a provocative test of GH secretion should be undertaken in accord with local endocrine practice and taking into consideration clinical indicators for use of insulin in epilepsy

Inpatient fasting
There is insufficient evidence to justify inpatient fasting without suggestive symptoms. However, hypoglycaemia is a particular risk in some circumstances, when less invasive screening using continuous glucose monitoring (CGM) is warranted. If CGM is not accessible or suitable, capillary plasma glucose monitoring during symptoms can be considered. We suggest that specialist endocrine opinion is sought during (a) evaluation of seizures and (b) prior to sedation, e.g. for neuroimaging, as some CGM devices are incompatible with MRI.

Patients with proven hypoglycaemia
The above evaluation of the GH axis should also be undertaken in patients with proven hypoglycaemia, in which case a dynamic test of the hypothalamic-pituitary adrenal axis should also be undertaken

Emergency management
This is conventional, depending on severity of hypoglycemia. It may range from infusion of intravenous glucose to provision of sugar-containing drinks or snacks. Glucagon injection may also be of use

Deficiency of free thyroid hormone levels
Given the importance of thyroid function for neurodevelopment, any deficiency of free thyroid hormone should be corrected with L-thyroxine. TSH is not a suitable marker of hormone replacement in central hypothyroidism. Monitoring should rely instead on free thyroxine concentrations. There is a lack of data on the use of T3 in this group

Long term management
Preventative measures depend on etiology of hypoglycaemia, and duration of fasting tolerated. In severe cases regular enteral or parenteral nutrition should be instituted with inclusion of overnight feeding, although modified starch preparations may progressively be used instead or as well in older patients

Deficiency of the adrenal axis
Should be corrected with cortisol. Note, however, that mild blunting of the axis may be a consequence rather than the cause of chronic hypoglycaemia due to resetting of counter-regulatory responses neurocognitive outcomes, 40,41 but this observation has not been tested specifically with respect to PROS. The average age of onset of epilepsy in association with PROS is also not known.
Seizures can occur anytime from the neonatal period to adulthood.

Recommendations for PROS
A referral to a pediatric neurologist/epileptologist is recommended if seizures occur in any individual with features of PROS. There is insufficient evidence for individuals with PROS who develop epilepsy, to recommend a specific antiepileptic drug (AED) or a management mainstay. Accordingly, we recommend following standard guidelines as outlined in Table 6(A) 42-44

| General discussion
There is a spectrum of neurodevelopmental impairment associated with PROS that includes ID, autism spectrum disorder (ASD), communication disorder, attention deficit hyperactivity disorder (ADHD), anxiety, other behavioral issues, and motor dysfunction.

| Recommendations for PROS
Neurodevelopmental concerns are of particular importance for individuals with PROS with brain involvement. We recommend a formal neuropsychological assessment in all individuals with PROS and with possible brain involvement (Table 1). In contrast, individuals without brain involvement do not always need neuropsychological assessment. Detailed recommendations are outlined in Table 4.

| General discussion
The underlying pathogenetic mechanism for this spectrum of disorders

| Recommendations for non-MCAP PROS
To date endocrinopathy in PROS has solely been reported in people with brain involvement. Although in principle physiologically inappropriate activation of PIK3CA in insulin-responsive tissues such as liver, muscle or adipose tissue could lower blood glucose, this has not been reported to date, even with extreme regional overgrowth of muscle and adipose tissue. 50 3.10 | Vascular anomalies and risk of thrombosis Section and PROS phenotype (see Table 1) Recommendation

Genetic testing
Klippel Trenaunay syndrome The confirmation of a PIK3CA pathogenic variant in these patients can be quite challenging and obtaining affected tissue at the time of surgery may offer the best chance. If use of inhibitors is going to be considered, demonstrating the mutation will be necessary, so planning and coordination will be critical

PROS
In any pregnancy where the foetus is known to have PROS, pregnancies may be complicated by polyhydramnios, prematurity, and difficult delivery in case of megalencephaly (foetal MRI) and/or breech presentation. The foetal management team should be made aware of these complications and arrangements for delivery made accordingly Critical care

PROS
Request cardiology review in all individuals after diagnosis and follow-up as appropriate in case of a diagnosis of cardiac arrhythmia In the presence of vascular anomalies assess the risk of thrombosis, hemorrhage and infections

PROS
The decision to use endoscopic third ventriculostomy, a ventricular diversion procedure (ventriculoperitoneal, ventriculoatrial etc.) vs. suboccipital decompression is beyond the scope of these recommendations. However, a relevant surgical consideration is that these procedures can carry an elevated risk of blood loss and postoperative vascular complications (e.g., stroke). While the incidence of intracranial vascular abnormalities in PROS is unknown, these patients can be at risk for vascular dysregulation. Therefore, surgical management decisions require a well-versed multispecialty team lead by a neurosurgeon to make appropriate recommendations that are individualized on a caseby-case basis MCAP with ventriculomegaly/hydrocephalus Historically, many children with MCAP and ventriculomegaly were empirically shunted in the absence of evidence for clear hydrocephalus. Given the rapidly progressive nature of megalencephaly in MCAP (with the head OFC crossing centiles especially early in life) combined with the co-occurrence of ventriculomegaly, close monitoring of the child is recommended including (a) close monitoring of OFC progression and trajectory, (b) watching for any signs/symptoms of increasing intracranial pressure/hydrocephalus, (c) close inspection of the brain MRI scan by an experienced neuroradiologist for any signs of increased ICP and/or obstructive ventriculomegaly. The decision of whether to intervene neurosurgically (whether by shunting or a third ventriculostomy) should be ideally made by a multidisciplinary team T A B L E 6 (Continued) PROS with hypoglycaemia and/or linear growth retardation A trial of GH may be indicated. In this case careful follow up of linear growth and trajectory of overgrowth should be undertaken, in case pathological growth is exacerbated by GH. Pragmatic delay of GH therapy until after 2 years of age has been suggested, avoiding the major period of brain growth. There exists a theoretical possibility that GH may promote tumourigenesis in the context of PROS in the longer term, as an "oncogenic" PIK3CA variant is already present. Reassuringly, to date no excess risk of neoplasia has been discerned in untreated individuals with MCAP. There are limited reports only of GH therapy in individuals with molecularly proven MCAP to date. It has generally been well tolerated with a single exception, and in at least four cases it proved effective with no discernible adverse effects. 51 Further evidence is required before a confident statement can be made about relative risks and benefits of GH therapy in GH-deficient PROS

PROS
Given the absence of systematic data, we recommend counseling the families on a case-bycase basis Oral/Dental management PROS with asymmetric facial overgrowth The most frequent problems are abnormal tooth eruption, dental crowding, gingival and periodontal pathologies and these should be treated as in the general population (see also Appendix)

Facial infiltrating lipomatosis
Orofacial presentation by Koutlas et al., 2021 10 MCAP B. Evidence score C: no evidence or consensus agreement/not currently specified as best practice (Methods and Figure 1) Approach to negative testing results

PROS
Currently, there is no evidence to direct how many times re-testing or re-biopsy should be considered where the a priori clinical probability is high. Each case must be assessed individually, taking into account previous experience with each methodology, obtaining new samples, or the need for molecular confirmation in the case of targeted therapies (see also Appendix) Somatic overgrowth (craniofacial, trunk-spine, limb)

MCAP
There are no specific growth charts for MCAP yet, which makes monitoring weight gain challenging

Endocrinopathies
The options here are rational but currently untested, and should be undertaken only under endocrine guidance on an experimental basis with due informed consent We also believe that our approach is complementary to conditionspecific guidelines. 62 66 We believe that outlining clinical management guidelines is essential in view of the on-going clinical trials and up-and-coming treatments for PROS. A standard of care can be useful for monitoring the effects of such trials or caring for non-responders.

CONFLICT OF INTEREST
LGB is a member of the Illumina ethics advisory board. RKS consults for Novartis. The rest of the authors declare no conflict of interest.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/cge.14027.

DATA AVAILABILITY STATEMENT
All data generated or analysed during this study are included in this published article. Genetic testing sampling: Our recommendations for MCAP include a two-step process that is to request biopsies of affected tissue only if the result is negative in blood or saliva samples. This however also depends on the family views and on whether they prefer to wait for different steps. A skin biopsy is an easy process, and some families may prefer a first step skin biopsy to have an answer more quickly. Ideally, this should be discussed with the families.

ORCID
Genetic testing results: These may have considerable importance for personal and medical decisions. Thus, the molecular report must be understood by non-specialist healthcare professionals. There is a high proportion of mutational hot spots in PIK3CA that facilitate reporting. It also should be accompanied by a clinical report explaining the implications and consequences of the results for that specific patient, where possible. In fact many laboratories do not always receive detailed clinical data linked to referrals so they may be limited in the ability to comment on the implications and consequences for each tested individual. As with other pathologies with a genetic cause, a negative test result does not rule out the clinical diagnosis.
Epilepsy and seizure: Though sometimes used interchangeably, the terms seizure and epilepsy are not the same. A seizure represents abnormal synchronous electrical activity in the brain, leading to neurological symptoms. In contrast, epilepsy is defined as having two or more unprovoked seizures occurring more than 24 hours apart or hav-