3. Pathological Characteristics and Etiopathogenesis
HIV-related DLBCL is histologically indistinguishable from DLBCL in the general population. However, some cases may show Hodgkin or Reed–Sternberg-like cells, raising the differential diagnosis with Hodgkin lymphoma (HL) or with polymorphic lymphoproliferative disorder. There are two main different morphological variants of DLBCL: the centroblastic variant (CB), composed of centroblasts with multiple nucleoli, and the immunoblastic variant (IB), composed of immunoblasts with a single and prominent nucleolus (
Table 2) [
12,
17]. The immunoblastic variant is considered to occur with more frequency in patients with more advanced HIV disease when compared with centroblastic variant [
39]. Moreover, it is known that HIV-related primary DLBCL of the CNS often presents with the immunoblastic type [
12]. In HIV-related DLBCL, there are more frequent plasmacytoid features compared to the general population [
39]. Although these cells present features of the immunoblastic stage of B-cell development, they also display a plasma-cell-related phenotype, expressing plasma cell surface markers, such as CD138, while mature B-cell markers (CD20, CD45) are often downregulated [
27]. The prognostic significance of the immunophenotypic characteristics is still unclear.
Table 2. DLBCL morphological variants in PWH: pathologic and immunophenotype markers, virologic co-infection, and genetic features. Adapted from Carbone et al. [
12,
17].
In the early 2000s, gene-expression profiling studies of DLBCL performed in the general population identified two transcriptional subgroups considering the supposed cell of origin (COO): germinal center B-cell subtype and activated B-cell subtype [
39]. The germinal center B-cell subtype of DLBCL corresponds to B-cells that are arrested at various stages of the germinal center transits, and the activated B-cell subtype of DLBCL seems to derive from germinal center B-cells evolving through plasma cell differentiation [
17,
39]. These DLBCL subgroups have different prognoses and may have relevance for treatment [
39]. However, the predictive power of the molecular classification is uncertain, and some data suggested that it is not clinically relevant for prognosis in HIV-related DLBCL [
40]. In contrast, most recent studies pointed out that HIV-related DLBCL cases with a non-germinal center phenotype tended to have a worse overall survival and progression-free survival than the cases with the germinal center phenotype [
41,
42].
Most of the evidence suggests that the etiopathogenesis of HIV-related NHL is a progressive multistep process, which involves viral and host factors and specific changes in the tumor clone [
43]. Human immunodeficiency virus is not known to have direct oncogenic effects, but HIV infection causes several indirect effects, which are implicated in lymphomagenesis. Recent evidence suggests that HIV may contribute to lymphomagenesis by acting directly on B-cells, as a critical microenvironmental modifier [
1]. In a systemic background of immunodeficiency (minimized by cART), B-cells receive an uncontrolled chronic activation through persistent antigenic stimulation, HIV CD40 ligand (CD40L), and HIV-encoded proteins, such as gp120, p17, and TAT [
1,
44]. Moreover, HIV infection leads to a loss of immune surveillance because of depletion of T lymphocytes, and reactivated oncogenic viral infection, such as Epstein–Barr virus (EBV), human herpesvirus 8 (HHV8), and chronic antigen stimulation mediated by other viral co-infections, including hepatitis B and C viruses [
6,
37]. In this way, co-infection with EBV or HHV8 contributes to certain subtypes of aggressive B-cell non-Hodgkin lymphoma development. Regarding DLBCL, EBV infection occurs in a proportion of cases, being more frequent in IB variants [
17]. HHV8 is causally associated with primary effusion lymphoma and, in fact, HHV8 infection of the cells is a diagnostic requirement [
45]. Aberrant B-cell activation and/or EBV infection may upregulate CXCR2 and IL-8 receptor, which are cellular receptors for p17 variants. CXCR2 upregulation by different p17 variants promotes B-cell expansions, increasing the probability of acquiring critical genetic alterations (involving
MYC,
BCL6 mutations, and other molecular events.) [
1]. Moreover, in EBV-infected B-cells, p17 variants may upregulate LMP-1, the main EBV carcinogenic protein, which would contribute to lymphomagenesis [
1].
EBV infection occurs in 90–100% of the cases with IB morphology and 25–30% of the DLBCL cases with CB morphology [
1,
17,
44]. In general, positivity for EBV is found in about 31% of the HIV-related DLBCL cases [
46]. After the primary infection, EBV spreads throughout lymphoid tissues and infects B-cells. Infected B-cells develop a primary cytotoxic T-cell response that would control the EBV infection and establish a reservoir of memory B-cells with latent viral expression. There are three known latency patterns (type I, II, and III) [
2]. As suggested in previous studies, EBV seems to be involved in different pathways depending on the DLBCL cell of origin subtype. Arvey et al. reported that 76% of germinal center B-cell cases are associated with latency type I (LMP1−, EBNA2−), 12% with latency type II (LMP1+, EBNA2−), and 12% with latency type III (LMP1+, EBNA2+). On the other hand, in the activated B-cell subtype, types II or III latency were both observed in 30% of the cases, and latency type I in 37% of the cases [
44].
EBV-positive HIV-related DLBCL presents a high expression of BLIMP1, repressing p53 transcription, conferring the ability to avoid apoptosis [
46]. Furthermore, HIV-related DLBCL with EBV infection is associated with CD30 expression [
46]. CD30 stimulates the activation of the NF-κB pathway, which is associated with cellular proliferation and carcinogenesis [
46]. Moreover, other genetic alterations involving
MYC,
BCL6, and
TP53 genes are frequently identified in HIV-related DLBCL [
2]. Additionally, aberrant somatic hypermutations involving
PIM1,
PAX5, and
RhoH/
TTF have been reported in about 50% of the cases [
2].
Due to the genetic and phenotypic heterogeneity of DLBCL, a DLBCL taxonomy was defined with regard to the genetic pattern. Schmitz et al. [
47] identified four prominent genetic subtypes in DLBCL, termed MCD (co-occurrence of MYD88L265P and CD79B mutations), BN2 (
BCL6 fusions and
NOTCH2 mutations), N1 (
NOTCH1 mutations), and EZB (
EZH2 mutations and
BCL2 translocations). Similarly, Chapuy et al. grouped DLBCL cases in five clusters according to their molecular characteristics [
48]. Recently, Wright et al. [
49] developed the LymphGen Classifier, unifying the two previous studies, and Lacy et al. [
50] also confirmed the existence of reproducible molecular subtypes of DLCBL defined by their profile genomic alterations, detected using a targeted sequencing panel applied to biopsy material. These classifications break DLBCL into different genetic subtypes that differ with respect to oncogenic pathway, gene-expression phenotype, tumor microenvironment, survival rates, and potential therapeutic targets [
49]. These studies have not been validated in HIV-associated DLBCL cohorts.
It is known that the tumor microenvironment and its interaction with viral components in HIV-related DLBCL play a crucial role in lymphomagenesis. HIV-related DLBCL is highly angiogenic and shows a markedly higher blood-vessel density than sporadic DLBCL cases [
51]. Some studies suggest that EBV infection may be related with these angiogenic properties [
51]. The duration of immunodeficiency, measured as the time since HIV-seroconversion, and the degree of chronic B-cell stimulation also have a role in lymphomagenesis and are measured mainly by means of two markers: raised serum immunoglobulin concentration and HIV p24 antigenemia [
43].
4. Clinical Characteristics, Treatment, and Prognosis
Regarding the clinical features, HIV-related DLBCL is characterized by advanced-stage disease at diagnosis [
36], the presence of B symptoms with more frequency than in the general population, and extranodal involvement at diagnosis, including bone marrow infiltration [
52] and leptomeningeal disease [
53]. Some studies performed in the cART era suggest that CNS involvement would have decreased, currently with a frequency similar to that of the general population [
54,
55]. Unusual extranodal locations are often reported, such as oral cavity, adrenal glands, kidney, lung, or bladder [
36]. As described in the Center for AIDS Research Network of Integrated Clinical Systems cohort of patients with HIV-related lymphoma between 1996 and 2010, the median CD4
+ T-cell count at diagnosis of HIV-related DLBCL is about 120/μL [
5].
Lymphoma may be the presenting manifestation of HIV infection, and all patients with aggressive B-cell lymphoma should be tested for HIV [
56]. Diagnosis of lymphoma should be based on an excision biopsy [
11]. Before starting the treatment, a staging procedure, including the same tests as the general population, should be performed. A basal
18Fluoro-2-deoxy-D-glucose positron emission tomography (
18F-FDG-PET) scanning at diagnosis improves the staging accuracy [
11,
57]. Trephine biopsy should be performed to rule out bone marrow involvement [
11]. HIV-related NHL is characterized by advanced-stage disease with an aggressive clinical course, and CNS involvement in HIV-related NHL occurs in about 13–20% of the cases [
55]; so, for most patients with HIV-related DLBCL, staging should also include evaluation for CNS involvement, with cytologic and flow cytometric analysis of the cerebrospinal fluid [
56,
58].
4.1. First-Line Treatment
In the pre-cART era, treatment with standard-dose chemotherapy induced high rates of toxicity and response rates were counterpoised by increased death due to opportunistic infections [
11]. In this context, low-dose chemotherapy and risk-adapted intensive chemotherapy regimens were administered to PWH [
12]. Outcomes were poor, with complete remission rates of about 50% and 5-year overall survival of about 28–47%, which was by far inferior to that observed in patients without HIV infection [
59,
60,
61,
62].
The introduction of cART led to an improved immune function in HIV patients, and the incorporation of infection prophylaxis and hematopoietic growth factors into treatment protocols allowed an increased use of conventional standard-dose chemotherapy regimens [
11]. The first-line treatments in PWH were the same used for the general population: CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisolone) [
55,
56] or infusional regimen, dose-adjusted (DA) EPOCH (etoposide, prednisone, vincristine, cyclophosphamide, and hydroxydaunorubicin) [
41,
63,
64]. In some centers, DA-EPOCH was preferred over CHOP, due to superior response rates observed when compared to historical data [
64,
65,
66]. However, a randomized study comparing CHOP vs. EPOCH in HIV patients has not been performed. Other centers have reported experience with infusional regimen of CDE (cyclophosphamide, doxorubicin, and etoposide) [
67], but there are no available data comparing CDE to CHOP. Guideline recommendations for HIV-related DLCBL are in the same line as those for DLCBL in HIV-negative patients, and CHOP is considered the standard first-line therapy [
68,
69,
70].
The addition of the anti-CD20 monoclonal antibody rituximab to standard chemotherapy regimens for B-cell NHL treatment has demonstrated a significant survival benefit in the general population. However, the use of rituximab in HIV-related DLBCL has been controversial [
71,
72]. In 2003, an AIDS Malignancy Consortium randomized phase 3 study with CHOP alone vs. CHOP with rituximab (
n = 150) was performed, and a trend was found for better outcomes associated with the use of rituximab [
73], but an increased frequency of infectious-related deaths in the rituximab group was detected (14% vs. 2%,
p < 0.001), particularly in those patients with CD4
+ lymphocyte counts <50/μL [
73]. Subsequent studies demonstrated that the combination of R-CHOP, R-CDE, R-EPOCH, or DA-EPOCH-R was beneficial (
Table 3), with an improvement in complete response rate (69–91%) [
41,
64,
71,
74,
75,
76], in 2-year progression-free survival (59–69%) [
64,
71,
74] and in 2-year overall survival rate (64–75%) [
58,
65,
68], with a lower infectious death rate (<10%) [
41,
64,
71,
74,
75,
76]. In a pooled analysis of 1546 patients, rituximab was associated with higher CR rate (odd ratio, 2.89;
p < 0.001), improved progression-free survival (PFS; hazard ratio 0.50;
p < 0.001), and OS (hazard ratio, 0.51;
p < 0.0001) [
77].
Table 3. Pivotal clinical trials with most commonly used chemotherapy regimens in HIV-related DLBCL. Extended from Carbone et al. [
17] and the references specified in the table.
Despite this evidence, the use of rituximab in patients with CD4
+ lymphocyte counts <50/μL is still controversial [
72]. In the cART era, treatment outcomes significantly improved for the patients with HIV-related NHL with CD4
+ lymphocyte counts <50/μL, and the 2-year OS increased to 65% from 16% in the pre-cART era [
41]. In this way, the observational study from Wyen et al. showed no association between rituximab use and infectious death risk in patients with a CD4
+ count lower than 100/μL [
79]. Nowadays, there is a consensus for rituximab use in all patients with HIV-related CD20-positive NHL, with special care in infectious prophylaxis in high-risk patients [
12,
17].
Several groups have pointed out that CDE or EPOCH infusional chemotherapy regimens [
41,
64] are associated with less tumor resistance, less cardiac toxicity, and the addition of a synergic effect with the rituximab combination [
80], and could be a better option in patients with HIV-related NHL. Analysis of the results of different trials from the AIDS Malignancy Consortium has suggested that in patients with HIV-related B-cell NHL, there is a higher efficacy for infusional R-EPOCH compared to R-CHOP bolus [
65]. Improved event-free survival (EFS; hazard ratio, 0.4;
p < 0.001) and overall survival (OS; hazard ratio, 0.38;
p < 0.01) were observed, and this difference was especially remarkable in patients with high-risk IPI [
65]. However, a randomized trial in immunocompetent patients with DLBCL showed that DA-EPOCH-R and R-CHOP were equally effective but with greater toxicity and complexity of the infusional therapy [
81]. It is controversial whether or not patients with a high-risk IPI could benefit from DA-EPOCH-R [
81].
4.2. Relapse/Refractory Lymphoma
High-dose chemotherapy platinum-based salvage regimens (e.g., R-DHAP, R-ICE, R-GDP, and R-EHAP) and subsequent autologous stem-cell transplantations (ASCTs) are strategies that have been used in relapsed/refractory HIV-related lymphoma [
82,
83,
84]. A multicentric study demonstrated that HIV infection has no impact on the long-term outcome of ASCT for lymphomas [
85]. The largest prospective study of ASCT in PWH was reported by the GICAT (Italian Cooperative Group on AIDS and Tumors), with 50 patients with relapsed/refractory HIV-related lymphoma, achieving a PFS and OS after 4 years of follow-up, of 75% and 76%, respectively, among the 27 patients who had undergone ASCT [
82]. Regarding the role of allogeneic hematopoietic stem-cell transplantation (HSCT), recent studies suggest that in the cART era it could be a feasible option for PWH with high-risk relapsed malignancies. However, interactions between immunosuppressive drugs and antiretroviral agents, the presence of some degree of graft-versus-host-disease, and the high incidence of infectious complications should be taken into account [
86]. Selected PWH with hematologic malignancies should be considered for allogeneic HSCT when indicated, in experienced centers [
86]. A few case reports have shown the feasibility of manufacturing CD19 CAR-T-cells in PWH and a successful outcome in some cases [
87,
88].
Unfortunately, most clinical trials for new agents and for cell therapy still exclude PWH [
89], and there is limited evidence for novel agents, such as polatuzumab vedotin, brentuximab vedotin, lenalidomide, and proteasome inhibitors in the HIV setting.
4.3. Antiretroviral Therapy during Chemotherapy
Maintenance or initiation of cART concomitantly with chemotherapy has demonstrated an improved complete response rate [
77,
90,
91,
92,
93] and immune recovery after chemotherapy [
94]. The risk of drug interactions associated with the use of strong cytochrome 3A4 inhibitors such as ritonavir- or cobicistat-based regimens has been overcome with new antiretroviral agents with fewer drug interactions, such as unboosted integrase strand-transfer inhibitors (INSTIs) [
17]. Therefore, most current guidelines recommend the use of cART during chemotherapy in PWH and lymphoma [
77,
95], preferably with unboosted INSTIs [
17].
Clinically significant interactions between chemotherapy regimens and cART have been reported, with a higher risk with regimens with ritonavir or cobicistat (boosters). These two agents are potent inhibitors of cytochrome P450 enzymes and could change the disposition of numerous drugs, leading to marked increases in drug exposure [
96,
97]. Of note, the use of boosters in PWH with lymphoma has been associated with higher probability of dose-reduction as well as with worse overall survival [
98,
99]. Similarly, the use of ritonavir increased the risk of adverse events in PWH receiving CHOP [
100,
101]. In addition, an increased autonomic neurotoxicity in patients receiving lopinavir/ritonavir with vincristine was described by Leveque et al. [
102].
Conversely to ritonavir or cobicistat, some non-nucleoside reverse transcriptase inhibitors (NNRTIs) such as nevirapine, efavirenz, and etravirine are moderate to potent cytochrome P450 inducers. Consequently, their use in PWH and lymphoma could potentially reduce the exposure, and thus the efficacy, of certain chemotherapy drugs [
103]. Rilpivirine and doravirine are second-generation NNRTIs that do not induce the P450 system, limiting their potential for interactions with chemotherapy [
104,
105]. As previously mentioned, cART with unboosted INSTIs may be particularly recommended for PWH and lymphoma due to their favorable efficacy, safety, and drug interactions profile. Raltegravir, dolutegravir, or bictegravir do not exert any inducer or inhibitory effect on P450 enzymes or drug transporters, minimizing their potential for drug interactions [
106,
107,
108].
The HIV capsid inhibitor lenacapavir and the HIV attachment inhibitor fostemsavir are two new drugs recently approved for the treatment of HIV infection. Although these drugs do not need to be boosted by ritonavir or cobicistat, they may exert some interactions that need to be considered. Lenacapavir is a moderate inhibitor of the isoenzyme CYP3A4 of the P450 system, and it may increase the exposure to other drugs that are metabolized by CYP3A4. Consequently, caution is advised when lenacapavir is co-administered with sensitive CYP3A4 substrates with a narrow therapeutic index [
109]. Similarly, temsavir (the active moiety of fostemsavir) inhibited the drug transporters OATP1B1/3 and BCRP in vitro, and administration of fostemsavir is expected to affect the pharmacokinetics of active substances that are substrates of OATP1B1/3 or BCRP. Therefore, close monitoring is recommended, and eventual dose modifications may be needed [
110].
Importantly, antiretroviral agents may also interact with other common drugs, such as omeprazole or other proton pump inhibitors, which reduces the bioavailability of rilpivirine, resulting in a decrease in the antiviral action [
104]. In the same way, antiacids containing divalent cations can reduce INTIs absorption [
111,
112]. An antiretroviral regimen can be changed before starting a chemotherapy regimen to avoid drug interactions and reduce toxicity in order to improve tolerability and adherence. Discontinuation of regimens with ritonavir or cobicistat is recommended, but the discontinuation of a single drug from the antiretroviral regimen must be avoided, because it may decrease the efficacy of cART and favors viral resistance. Modifications in cART regimens should be consulted with an HIV specialist and interdisciplinary evaluation is mandatory in order to decide the best treatment for both the hematologic malignancy and the HIV infection [
113].
4.4. Additional Measures and Supportive Care
Infection prophylaxis must be performed in PWH receiving cancer treatment [
11,
12,
114,
115].
Pneumocystis jiroveci prophylaxis is indicated in PWH receiving chemoradiotherapy because chemotherapy can markedly reduce CD4
+ T-cell counts, even in patients on cART.
P. jiroveci prophylaxis is usually performed with oral cotrimoxazole three times per week (800 mg trimethoprim and 160 mg sulfamethoxazole). Aerosolized pentamidine 300 mg once a month, dapsone 100 mg daily, or atovaquone 1500 mg daily are alternative options. Prophylaxis for
Mycobacterium avium complex should be considered for patients with CD4
+ counts <50–100/μL, with oral azithromycin 1200–1250 mg per week or rifabutin 300 mg daily. Primary prophylaxis of the chemotherapy-related neutropenia with a granulocyte colony-stimulating factor must be performed, starting 48–72 h after chemotherapy. In patients with neutrophil count of <100/μL or neutropenia lasting more than 7 days, antibacterial prophylaxis with levofloxacin could be considered, although it could favor the emergence of multiresistant bacteria. PWH with low CD4
+ T-cell count are at an increased risk of fungal infections, particularly candidiasis and cryptococcosis. Therefore, fluconazole is recommended for PWH receiving chemotherapy. PWH treated with intensive chemotherapy should receive acyclovir or valacyclovir for herpes simplex and varicella prophylaxis. Monitoring for cytomegalovirus infection must be performed. Patients with hepatitis B virus infection or positive anti-HBc receiving chemotherapy should be treated with antihepatitis B virus agents. Patients with hepatitis C virus infection must receive antiviral treatment even if the fibrosis stage is detected [
116,
117]. Moreover, all PWH with hematological malignancies should receive annual vaccination against SARS-CoV-2 and influenza [
116,
117,
118].
4.5. Prognosis
During the pre-cART era, HIV-related NHL outcomes depended specially on HIV-related factors, such as a poor bone marrow reserve, CD4
+-cell count <100/μL, prior AIDS-defining illness, HIV viral load, or the development of an opportunistic infection [
36]. In the Straus et al. study, an index was developed which included CD4
+-cell count, age higher than 35 years, stage III or IV disease, history of intravenous drugs, and elevated serum lactate dehydrogenase (LDH) [
119]. In the same vein, GICAT (Gruppo Italiano Cooperativo AIDS e Tumori) and GELA (Groupe d’Etude des Lymphomes de l’Adulte) groups performed an index combining three independent risk factors: ECOG (Eastern Cooperative Oncology Group), performance status of 2 or more, prior AIDS defining illness, and CD4
+-cell count <100/μL [
120].
In the pre-cART era, outcomes were poorer and highly influenced by hematological toxicity (complete remission rates of about 50% and 5-year OS of about 28–48%), with a huge difference compared to HIV-negative patients [
12]. In the cART era, the paradigm of therapy has significantly changed in PWH with lymphomas. In addition to better control of viral load, better management of the chemotherapy regimens and better supportive care is carried out, and survival is now reaching that of patients with non-related-HIV NHL [
12]. Therefore, currently HIV-related DLBCL prognosis is determined by lymphoma features: international prognostic score (IPI), revised-IPI [
121,
122], or age-adjusted IPI, rather than HIV-specific factors [
17]. Some authors have focused on the need for refinement of the aa-IPI score in patients with HIV-related NHL, due to the HIV infection being a competing risk that may influence prognosis. In 2014, Barta et al. developed a new prognostic index called ARL (AIDS-related lymphoma)-IPI, which consists of three components: aa-IPI, number of involved extranodal sites, and an HIV score that incorporates baseline CD4
+ count, HIV viral load, and prior history of AIDS [
123].
Despite the prognosis predicting power of the IPI being proved in HIV-related DLBCL [
123,
124,
125,
126,
127], new scores with a more conscious predictive ability have been developed (National Cancer Comprehensive Network IPI (NCCN-IPI) [
128,
129], GELTAMO-IPI [
130], and another new score which includes data from peripheral blood count [
131]), including new variables such as beta2-microglobulin and/or lymphocyte and monocyte count, is still not validated in the HIV-related DLBCL.